Reviewer #2 (Public Review):

The fruit fly represents an excellent model organism to analyze in detail how odor stimuli are processed by brain circuits at different levels of complexity. Whereas much insight has been gained in recent years about peripheral processes at the level of sensory neurons and the primary circuits within the antennal lobes, the roles and interplay of the many and diverse neurons at higher brain levels such as the lateral horn remain less clear. The authors have asked how glutamatergic neurons of the lateral horn respond to odors, i.e., whether they selectively encode odor identity, from which projection neurons they receive input, and whether a correlation between spatial response "pattern" and behavioral valence exists. Using state-of-the-art methods they demonstrate that, indeed, odor representations across these neurons are odor-selective, and that the responses correlate with the atractiveness of the odor. Moreover, inhibitory and excitatory inputs from two antagonistically acting populations of projection neurons appear to shape the response patterns and contribute to the generation of such a valence code.

The experiments are excellently documented, the approaches are convincing and modern, the statistical analyses are sound and the manuscript is very clearly written. The overall conclusions are convincing and add an important point to our understanding of the olfactory pathway. I congratulate the authors to this excellent study and their interesting findings.

Reviewer #3 (Public Review):

The lateral horn comprises multiple types of neurons, including a large fraction of glutamatergic neurons. This work investigates by functional imaging the general response properties of these glutamatergic neurons to odours of different valence. Complementing previous work, they find a certain degree of spatial stereotypy of response to odours and a correlation between amplitude of response of LH glutamatergic neurons and odour valence, which is not simply derived by the uPN input pattern. This is a solid piece of work improving our understanding of odour representation. The specific role of uPNs and iPNS in defining odour valence could be strengthened by an increased number of odours used.

Reviewer #1 (Public Review):

Here, the authors used multiple F1 crosses and the resulting embryonic fibroblasts to perform molecular profiling with ATAC-seq and a combination of ChIP-seq, Hi-ChIP, and CUT&RUN on multiple modified histones and transcription factors proteins. The resulting data are a good resource for quantifying allelic bias in protein-DNA binding and chromatin accessibility.

The authors claim there's "enrichment of SNPs/indels within a 150 bp window" in enhancers (Fig. 2H), but this enrichment looks quite middling. Can they quantify the level of enrichment and is it significant?

Reviewer #2 (Public Review):

Yang et al. apply genome-wide profiling of gene expression, accessibility, and ChIP-seq for CTCF and FOS occupancy, and H3K27ac, H3K4me1, H3K4me2, and H3K4me3 to a series of F1 hybrid mice derived from 9 divergent strains/species. The authors find that loss of AP-1 binding coincides with loss of H3K4me1 along with H3K27ac. They argue that most observed changes in accessibility or occupancy derive from cis effects rather than trans. The authors identify that while AP-1 binding does not rely on co-binding by TEAD, TEAD occupancy frequently is lost when a nearby AP-1 site is genetically perturbed. This is an interesting investigation of the dependencies between TFs binding nearby at the same accessible site, and will likely prove a useful resource for the field.

Reviewer #1 (Public Review):

The strength of this paper is identifying a novel factor that might be involved in the generation and/or maintenance of NK cells. To date, the full molecular network guiding the development, maturation and maintenance of NK cells has not been fully defined.

While the approach aims to dissect out this network and identified a number of differentially expressed genes, relatively little refinement or validation of that network has been performed. Thus, the study is somewhat descriptive and awaits a more detailed analysis and confirmation of interacting molecular partners.

Reviewer #2 (Public Review):

Transcription factors represent key regulators of NK cell development and function. This study compared two mouse strains with or without Bach2 expression in NK cells for NK cell development, maturation and function. The phenotype of wild-type and Bach2-deficient NK cells were investigated by flow cytometry and transcriptomic profiling. The effector function of NK cells was evaluated in a mouse model of NK cell-mediated protection of tumor metastasis in the lung. Authors conclude that Bach2 expression in NK cells negatively regulates NK cell maturation and effector function.

The results support the key conclusion. However, some experimental information was missing and additional analysis of NK cell function in the tumor model can improve the knowledge gained from this study.

Reviewer #3 (Public Review):

Motivated by previously described roles for BACH2 in T cells and B cells, Li et al examine the role of BACH2 in steady-state NK cell differentiation. They first examine the kinetics of BACH2 expression during NK cell development and differentiation using a reporter mouse line. After BACH2 initially decreases in expression once common lymphoid progenitors commit to the NK lineage, BACH2 expression is regained within mature NK cells across multiple anatomical sites. Interestingly, within mature NK cells, BACH2 expression is lost within terminally differentiated NK cells, which is analogous to the BACH2 expression patterns seen in exhausted CD8+ T cells. Subsequent phenotyping of mice with conditional BACH2 deletion within NK cells revealed an accumulation of more effector-like, terminally differentiated NK cells at the expense of less differentiated populations. Challenge of these conditional knock-out mice using a lung melanoma metastasis model known to be highly NK cell sensitive revealed superior tumour control. Collectively, this suggests that BACH2 restrains differentiation of terminally differentiated effector NK cell subsets and that targeting BACH2 in NK cells could represent a novel strategy for cancer immunotherapy.

Strengths

This is the first paper to comprehensively describe BACH2 expression patterns across NK cell development and differentiation, and the first paper to describe a function for BACH2 in NK cells. There are many interesting parallels between BACH2 expression patterns and function in NK cells, and those previously reported in CD8+ T cells, suggesting a conserved function for BACH2 in both contexts. Nevertheless, there are also some interesting differences. Of note, one of the key mechanisms by which BACH2 controls exhausted CD8+ T cell differentiation is through repression of Blimp-1 and BATF, however, neither factor appears to be up-regulated after BACH2 loss in NK cells suggesting key mechanistic differences. Finally, the authors provide interesting data suggesting that the boosted effector function of BACH2 knock-out NK cells leads to elevated tumour control, which represents a novel approach to boosting NK-dependent tumour control.

Weaknesses

There are a few key weaknesses that limit the impact of the study. In particular, the mechanistic explanation for how BACH2 mediates the observed phenotypes is very limited. Furthermore, published work on BACH2 biology rules out some of the authors' claims. For example, the authors speculate that BACH2 operates similarly in TCF1+ exhausted CD8+ T cells and NK cells, as in both cases knock-out cells lose TCF1 expression. However, as noted above, the effects of BACH2 on TCF1 in CD8s are indirect and due to both repression of Blimp-1 and BATF expression, and antagonism of RUNX3 and BATF binding site accessibility. Given that Blimp-1 and BATF expression is unchanged in BACH2 knock-out NK cells, this argues that there are significant mechanistic differences between how BACH2 operates in CD8+ T cells versus NK cells. Moreover, unlike CD8+ T cell phenotypes, many of the phenotypic changes in BACH2 knock-out NK cells are very subtle again hinting at underlying differences in the mechanism.

The authors suggest that the changes in NK cell subset composition are responsible for the enhanced control of melanoma lung metastases in knock-out mice, but alternative explanations are not explored. For example, it is possible that enhanced NK cell homing to the lung could explain these phenotypes. Data examining knock-out NK cell effector functions (eg. ex vivo killing capacity and cytokine production) are also not included, so it remains unclear whether the subset changes in knock-out mice lead to a meaningful difference in effector capacity.

Finally, many of the figures show representative data without pooled data points and statistics, making it difficult to evaluate how consistent the reported trends are. Additionally, some experimental groups are small and would benefit from more mice, particularly for key experiments (eg. the melanoma metastasis data).

Reviewer #1 (Public Review):

Two important goals in evolutionary biology are (i) to understand why different species exhibit different levels of genetic diversity and (ii) in each species, what is the evolutionary nature of genetic variants. Are genetic variants mostly neutral, deleterious, or advantageous? In their study, Stolyarova et al. looked at one of the most polymorphic species known, the fungus Schizophyllum commune. They found that in this hyperpolymorphic species, the evolutionary forces that govern and structure genetic variation can be very different compared to less polymorphic species, including humans and flies. Specifically, the authors find that a process known as positive epistasis is quantitatively abundant among genetic variants that alter proteins in S. commune. Positive epistasis happens when a combination of multiple genetic variants is advantageous for the individuals that carry them, even though each isolated variant in the combination is not advantageous or even detrimental on its own. The authors explain that this happens frequently in their hyperpolymorphic species because the very high polymorphism level makes it very likely that the genetic variants will by chance occur together in the same individuals. In less polymorphic species, the variants that are advantageous in combination may have to wait for each other to occur for too long, for the combination to ever happen often enough in the first place.

Overall I had a great time reading the manuscript, and I feel that my understanding of evolution has been advanced on a fundamental level after reading it. However part of the reason why I enjoyed it was having to fill the gaps, answer the riddles left unanswered in the story by the authors.

Strengths:

1) The model, both extremely polymorphic and amenable to haploid cultures, is ideal to address the questions asked.

2) The study potentially represents a very important conceptual advance on the way to better understand genetic variation in general.

3) The interpretations made by the authors of their data are likely the correct ones to make, even though more definitive answers will likely only come from the sequencing of a much larger number of haplotypes, which cannot reasonably be asked of the authors at this point.

Weaknesses:

1) The manuscript does not provide enough information to judge if the synonymous controls that are compared to the nonsynonymous variants are fully adequate. Specifically, I have one concern that the Site Frequency Spectrum (SFS) of the synonymous variants at MAF>0.05 may be very different compared to the SFS of nonsynonymous variants at MAF>0.05. I focus on this because the authors mention page 5 line 3: "The excess of LDnonsyn over LDsyn corresponds to the attraction between rare alleles at nonsynonymous sites". First, it is unclear from this or from the figures at this point in the manuscript what the authors mean by rare alleles, among those alleles at MAF>0.05. This needs to be detailed quantitatively much more carefully. Second, and most importantly, this raises the question of whether or not the synonymous controls have a SFS with many less rare (but with MAF>0.05) alleles, as one may expect if they are under less purifying selection than nonsynonymous variants. This then raises the question of whether or not the synonymous control conducted by the author is adequate, or if the authors need to explicitly match the synonymous control in terms of SFS for MAF>0.05 in addition to the distance matching already done.

2) The manuscript is far too succinct on several occasions, where observations or interpretations need to be much more detailed and explained.

Reviewer #2 (Public Review):

Stolyarova et al. used a highly polymorphic species, Schizophyllum commune, to explore patterns of LD between nonsynonymous and synonymous mutations within protein-coding genes. LD is informative about interference and interactions between selected loci, with compensatory mutations expected to be in strong positive LD. The benefit of studying this fungal species with large diversity (with pi > 0.1) is that populations are able to explore relatively large regions of the fitness landscape, and chances increase that sets of epistatically interacting mutations segregate at the same time.

This study finds strong positive LD between pairs of nonsynonymous mutations within, but not between genes, compared to pairs synonymous variants. Further, the authors show that high LD is prevalent among pairs of mutations at amino acid sites that interact within the protein. This result is consistent with pairs or sets of compensatory nonsynonymous mutations cosegregating within protein-coding genes.

The conclusions of this paper are largely supported by the data, with some caveats, listed below.

1. With such large pairwise diversity, there are bound to be many deleterious variants segregating at once, and the large levels of interference between them will make selection much less efficient at purging deleterious variants. While the authors argue that balancing selection is needed to account for patterns of haplotype variation they see, widespread balancing selection may not be required in this setting, and soft or partial selective sweeps (either on single mutations or sets of mutations) can also lead to patterns of diversity where a small number of haplotypes are each at appreciable frequency.

There is also a tension between arguing that balancing selection is widespread and that shared SNPs across populations are expected to arise through recurrent mutation, as balancing selection is known to preserve haplotypes over long evolutionary times. In that section of the discussion especially, I had difficulty following the logic, and some statements are presented more definitively than might be warranted.

2. The validations through simulation are somewhat meagre, and I am not convinced that the simulations cover the appropriate parameter regimes. With a population size of 1000, this represents a severe down-scaling of population size and up-scaling of mutation, selection, and recombination rates (if > 0), and it's unclear if such aggressive scaling puts the simulations in an interference/interaction regime far from the true populations. A selection coefficient of -0.01 also implies 2Ns = -20, whereas Hill-Robertson interference is most pronounced between mutations with 2Ns ~ -1.

3. Large portions of the genome (8.4 and 15.9%, depending on the population) are covered by haploblocks, which are originally detected as genomic windows with elevated LD among SNPs. It's therefore unsurprising that haploblocks identified as high-LD outliers have elevated LD compared to other regions of the genome, and the discussion about the importance of haploblocks seemed a bit circular.

4. Finally, the authors observe a positive correlation between Pn/Ps and LD between both synonymous and nonsynonymous mutations. This result is intriguing and should be discussed, but the authors do not comment on this result in the Discussion.

Reviewer #1 (Public Review):

The manuscript by Liu et al investigates how MRI can be used to detect the earliest stages of CNS infections and how MRI can also be used as a surrogate readout for treatment efficacy. Authors demonstrate convincingly that microbleeds, as evidenced by unusual dark spots in the brain of mice infected with a virus that infects the brain, occurred at the earliest stages of viral infection. Authors also convincingly demonstrate that the infusion of virus-specific immune cells, when delivered at the right time and at the right dose, could reduce these microbleeds. Importantly, authors showed that the wrong dose could be detrimental.

The authors cast this study as a method for improving research and discovery in immunotherapy context and the study is convincing in its conclusions regarding imaging microbleeds and the immunotherapy tested herein. While authors do not directly suggest so, these findings extend the significance of this work beyond research and development of immunotherapies by providing a potential early detection mechanism for viral infection in the brain. This may be feasible as the MRI methodologies for detecting these phenomena are generally translatable to clinical imaging scenarios, though the imaging resolution may not.

Weaknesses in the report revolves around the value of and the ability to image magnetically labeled T cells in the presence of microbleeds.

1) Authors developed a magnetic particle coated with fluorescent molecules and antibodies specific for CD8+ T cells. They labeled these T cells with particles for detection by MRI. They then wanted to follow the accumulation of these cells in the brain following infusion and viral infection by performing MRI using parameters that amplify the signal of the attached label. The rationale for these experiments was to determine if immune cell infiltration preceded vascular compromise. This suggests the expectation for active chemotactic migration or other signaled accumulation rather than leakage. When authors tested their magnetically labeled T cells for functional impairment due to the presence of attached magnetic particles, they did not test for deficits to migratory capabilities, such as in standard transwell migration assays. Others have shown (see https://doi.org/10.1038/nm.2198 for example) that T cell migration is very sensitive to the type of attached nanoparticle as well as the surface coverage. Perhaps authors should temper their claims that magnetically labeling of T cells does not alter T cell function without at least an assay of this critical function. Further, the fluorescence microscopy shown in Figure 7D is of insufficient resolution to claim that MPIOs are inside cells. Electron microscopy should be used to determine this.

2) Regarding the use of imaging the accumulation of magnetically labeled T cells, authors show evidence that magnetically labeled T cells accumulate in areas of the brain that as yet do not present with microbleeds but do have the histological hallmarks of vascular inflammation. This corroboration is intriguing but only provable with a serial imaging study in the same animal, which was not performed. Authors are also encouraged to report on the frequency in which a magnetically labeled T cell was present in a pre-vascular compromised inflammatory environment. The bulk of the results on imaging magnetically labeled T cells essentially show that the accumulation of magnetically labeled T cells enhances the ability to detect microbleeeds that otherwise were perhaps too small to detect (Sup Fig 8). Given the lack of data supporting the retained migratory capacity of magnetically labeled T cells, one wonders then, whether magnetically labeled T cells are indeed trafficking to the brain or are passively arriving in the brain, and might some vascular magnetic particle accumulate in an early inflammation or leak into the microbleed on its own and similarly enhance the ability to detect the otherwise undetectable microbleed. A series of controls would be useful to answer these questions, perhaps testing the administration of magnetic particles alone, and/or magnetically labeled non-CD8+ T cells. Authors are also encouraged to report on the frequency in which a magnetically labeled T cell was present in a pre-vascular compromised inflammatory environment versus in the microbleed, as measured by MRI and histology.

Reviewer #3 (Public Review):

The manuscript by Liu et al., aims to develop a novel MRI-based approach to monitor virus specific CD8+ T cells and their relationship to cerebrovascular pathology in living brains. Using a mouse model of VSV brain infection, they show that MRI approaches can be used to identify microbleeds in the brain, these microbleeds occur independent of immune cell influx, and that the transfer of low numbers of virus specific CD8+ T cells can reduce cerebrovascular bleeding.

Reviewer #2 (Public Review):

Imaging based methods to detect early vascular damage and to better understand the relationship to invading pathogens and immune cell infiltrates is relevant to both identify sites of pathology and develop targeted therapies to control infections while minimizing pathological events. The paper in its current form uses high resolution MRI combined with histological analysis and T cell labeling to better define the relationship between microbleeds indicative of cerebrovascular damage, viral replication and CD8 T cell accumulation. Several interrelated aims address whether the detection of microbleeds can serve as a neuroimaging marker for infection and to what extent microbleeds are associated with direct virus infection or infiltrating immune cells.<br /> Central nervous system (CNS) vascular bleeding has been associated with virus infections, but the underlying mechanisms are poorly defined. This paper overall demonstrates the suitability of high resolution MRI to detect both micro bleeds and preferential sites of T cell inflammation throughout the CNS in small animal models. MPIO labeling and high resolution MRI with the sensitivity to detect single cells is applied to nonphagocytic lymphocytes and should be applicable to track accumulation of other peripheral immune cells of interest in a variety of neurological disease models. The technology should be of wide interest to the field of neuroinflammation. The data support the aims and are well documented. It is also acknowledged that more research is required to distinguish causation of pathogenic signals. The ability to non invasively monitor initial insults associated with infections as well as evaluate therapies would advance both detection and treatment outcomes of brain infections, where the causative agents are often unknown. It remains to be determined if results apply uniquely to VSV or is similarities are noted in other models. It will also be of interest to pursue the role of glia activation in promoting microbleeds in the absence of peripheral leukocytes.<br /> Strengths:

• This work takes advantage of high resolution MRI to detect areas of cerebrovascular breakdown following viral infection with VSV. VSV is known to enter the brain when administered intranasally and provides a good model to study entry of viruses from the nasal cavity into the olfactory bulb and other brain areas.<br /> • Segregating events leading to vascular breakdown has been difficult. Results from MRI and histology show that microbleeds can be directly associated with infection of vascular cells, even when entry of peripheral immune cells is blocked. On the other hand, transfer of activated CD8 T cells at onset and peak infection reduced levels of infection and coincidently microbleeds. This finding is relevant as immune cells, including anti-viral cytolytic CD8 T cells, are commonly associated with vascular damage.<br /> • The measurement of microbleed numbers and volume with T2* is appropriate to demonstrate affected regional sites, changes over time, and differences across treatment. Hyperintensities evident by MRI allow precise localization of affected areas for subsequent more refined histological analysis.<br /> • An exciting novel aspect of the paper is the development of chemically modified MPIO particles to improve internalization and labelling efficiency of T cells ex vivo for subsequent transfer and MRI detection in mice. The technique is well described and extensive data showing efficacy of labeling and no effects on select T cell functions are included. MRI combined with histological analysis revealed hypointensities were attributed to colocalization of microbleeds and labeled CD8 T cells in some areas, but labeled CD8 T cells alone in others.

Weaknesses:<br /> • Individuals with systemic infections or other underlying condition may have microbleeds due to inflammation or hypertension. The etiology of microbleeds is thus not necessarily tied to CNS infections. Investigation of potential cerebrovascular microbleeds following systemic or respiratory infections not affecting the CNS may shed light on this possibility which may also provide alternative interpretation of neurological symptoms associated with on CNS invasive infections.<br /> • Representative colocalization of virus infected endothelial cells with red blood cells (RBCs) is shown in Fig 4. However, a more quantitative assessment indicating how many areas or hypointensities were evaluated for virus-localization with RBCs, and how many of these revealed colocalization versus virus or RBC only would strengthen interpretation.<br /> • A limitation clearly acknowledged by the authors is that hypointensity spots detected by MRI cannot distinguish microbeads from MPIO-labeled T cells.

Reviewer #1 (Public Review): 

In comparison to closely related archaic genomes (i.e., Neanderthal and Denisovan), modern human lineage has an elevated rate of nonsynonymous substitutions in some spindle protein genes (first reported in Prüfer et al. 2014). Following up on this interesting observation, Peyrégne et al performed a detailed study on the human lineage substitutions using both present-day and archaic genomes. In particular, they reported the back introgression of the kinetochore scaffold 1 (KNL1) gene. Using the genetic divergence and segment length, the authors inferred that KNL1 first introgressed from an ancestral modern human lineage to late European Neanderthals, then introgressed back to out-of-Africa modern humans. Surprisingly, they find no evidence for adaptive introgression of KNL1 in Neanderthals, despite the substitutions likely being adaptive in humans, and the Neanderthal copy of KNL1 having been purged from modern humans. Their nonadaptive conclusion is drawn upon the high frequency of other human variants in late European Neanderthals. However, reconcilation with the estimated 3% human to Neanderthal introgression by Hubisz et al. 2020 might be needed. The missense substitutions in KNL1, and the differences with the archaic copies, are worth following up in functional studies. Overall, the study nicely uses various population genetic approaches to understand the evolution of these spindle genes. My main concerns are about the robustness of the statistics because of the small sample size of Neanderthals and the low coverage. In particular, it is important to know whether these spindle protein genes are truly outliers in the genome-wide scan, and whether these results are robust to different variant calling protocols for the archaic genomes.

Reviewer #2 (Public Review): 

The authors set out to study spindle protein genes with missense changes that are fixed in modern humans relative to Neandertals and to reconstruct their evolutionary history. The authors show evidence for positive selection, particularly in the gene SPAG5, and they demonstrate that modern human fixed derived haplotypes at the gene KNL1 were transferred by gene flow from modern humans to early Neandertals and then again back from late Neandertals to modern humans. 

The main argument for the gene-flow hypothesis at KNL1 rests on the timing of mutations and gene flow events by investigating haplotype lengths and divergence within modern humans and between modern humans and Neandertal genomes. This part of the analysis seems technically sound and indeed suggests that a simple model of an ancient separation between Neandertals and modern humans and recent (~50kya) contact does not fit the data, i.e. a second gene flow event at KNL1 around 200kya has to be invoked. The analysis relies strongly on knowing the true mutation rate and the recombination rate in this genomic region. However, by using different datasets to estimate these parameters locally in the genome the authors show that their analysis is at least robust and consistent with the current best estimates of these two parameters in this region. I thus think their conclusion is well justified. 

The analysis reveals two other interesting observations: 1) The two missense changes in KNL1 predate the divergence of Neandertals and modern humans. However, for some reason Neandertals did not inherit these mutations from the ancestral population, only modern humans did. 2) Only derived KNL1 haplotypes that were back-introduced into modern humans from Neandertals persisted until the present day. The ancestral haplotypes were either not transferred or were selected against. I think both observations are only consistent with purifying selection acting at certain points in time in certain populations. However, the current analysis does not focus on this aspect. 

Further, the missense mutation in SPAG5, which seems to have been under positive selection in modern humans and which is not found in Neandertals, is dated to be older than the split between Neandertals and modern humans (Fig. 2C). This suggests that the mutation was not immediately under selection and that it was potentially under purifying selection in Neandertals. 

In general, the study provides very careful analysis and excellent usage of methods for estimating the timing of mutations and introgression events. It suggests that there had been contact between modern humans and Neandertals around 200kya, which is very relevant for understanding human evolutionary history. Further, it suggests strong selection of nonsynonymous mutations at spindle genes. This result will hopefully stimulate further research and functional characterization of these mutations in the future and lead to a better understanding of when and why these mutations had been selected.

Reviewer #4 (Public Review):

Smalley et al. investigated the evolutionary dynamics of the quorum-sensing regulon in 5 replicate populations of Pseudomonas aeruginosa (PA), in an environment maintaining selection for both cooperative and private QS-controlled traits. Building on previous studies illustrating differences in regulon size and content across PA isolates, the authors hypothesized that evolution over 1000 generations in a single defined environment would result in reduction in the size of the QS-controlled regulon. In agreement with this hypothesis, transcriptomic profiling of 2 of the 5 replicate lines indicated substantial (45% plus) reductions in QS regulon size (transcriptome data on the other 3 replicates were not shown). Molecular characterization of a gene lost from the QS regulon (pqsA) indicates this loss in some replicates was due to a mutation in it's QS-controlled transcriptional activator pqsR.

Strengths:

The primary strength of this paper is that it provides a direct attack on a series of basic questions of QS evolution - Can constant environments lead to the loss of genes from the QS regulon? On what timescale are they lost? What are the molecular mechanisms of loss? More broadly, the study sheds light on basic questions of the evolution of behavioral responses in bacteria.

More specifically, the study provides a clear answer on the question of timescale (substantial change within 1000 generations), which has implications for inferring recent environmental history from regulon profiling. On the question of molecular mechanisms, the results are more anecdotal but do extend our understanding of how QS systems can be rewired during evolution.

Methodologically, the study showcases and benchmarks an experimental design to query QS-regulons without constructing genetic QS knockouts, using instead a QS-quenching acylase treatment to turn QS regulons off, and QS signal supplementation to ensure QS regulons are on. Benchmarking this study on a prior QS regulon experiment is an important step to improve confidence in the results.

Weaknesses:

Limiting detailed mechanistic characterization to smaller number of evolutionary replicates is a standard step, yet it is important to make sure there is a clear rationale for the choice made, with careful consideration of how these choices could cause biased results. In the present study, 2 replicates from 5 were selected for detailed characterization, and I flag at this point that the rationale is not entirely evident. 2 replicates were discarded from consideration as they evolved to overcome the regulatory trap posed by the combination of a protein + adenosine environment with QS-control of protein (cooperative) + adenosine (private) metabolism. This was likely due to a regulatory rewiring to uncouple adenosine from QS-control, allowing the strain to then 'cheat' on protein degradation via an additional QS regulon loss. This sounds very much like a multi-step reduction in QS regulon, and it is not clear why this interesting path was discarded.

The hypothesis as stated is already well supported by the common experimental evolution and clinical observation of rapid lasR mutant expansion. In mitigation of this point, the study takes steps to reduce selection for lasR mutants (via adenosine supplementation), and also seeks to directly characterize regulon size and content, which is a valuable and rarely-taken step.

Likely impact:

The study holds the potential for broad impact on research into regulatory control of behavior in bacteria, providing a key benchmark study profiling how QS regulons can rapidly shrink even when QS-controlled genes are under continued positive selection. Future studies will likely refine and build out the generality of this result, but examining evolutionary dynamics under different environments (e.g. requiring activity of different sets of QS genes, or fluctuating requirements), mechanisms of re-wiring for other genes and in different species and strains, and assessing convergence across replicates.

Reviewer #3 (Public Review):

Pseudomonas aeruginosa is an important opportunistic pathogen causing life-threatening infections. Many of the virulence factor of this pathogen are regulated via chemical communication systems, referred to as quorum-sensing (QS) systems. There is great interest to understand QS evolution, as evolutionary changes affect the virulence potential of this pathogen. There are many studies that examined QS evolution in vitro and in hosts, and there is a consensus in the field that QS is under selection and that many QS-regulated traits are lost in long-term experiments and chronic infections.

While previous studies typically focussed on specific QS-phenotypes, the current paper applies a transcriptomics approach, which allows to identify global changes in QS-regulons. This is an interesting and novel approach. However, the study also comes with two big limitations. First, sample size equals N=2, which is so small that no general conclusions can be made. Even more problematic is that there were initially 5 replicates, and two of those replicates clearly took a different evolutionary path. They were deliberately excluded, which further compromises any claims on generality. Second, there was only one treatment (growth in CAB medium). Thus, we gain no information on how QS-regulon size and composition evolves across environments. Is QS-shrinking a general phenomenon or is it a CAB-specific pattern? We simply cannot tell.

Reviewer #1 (Public Review):

Pseudomonas aeruginosa is currently widely used as good model bacterium to study the evolution and maintenance of cooperative behaviors in bacteria. This bacterium uses quorum sensing to regulate hundreds of different genes involved in group behaviors and its attractive because it is easy to manipulate, and because it is relevant in many environmental and clinical settings.

In the manuscript "Evolution of the quorum sensing regulon in cooperating populations of Pseudomonas aeruginosa" the authors used experimental evolution to try to understand why some strains of Pseudomonas aeruginosa control hundreds of genes through quorum sensing while others have a much smaller number of genes regulated by quorum sensing. Here they evolved different P. aeruginosa populations in an environment where most of the of the cooperative processes regulated quorum sensing are not needed for growth. Evolution under this condition resulted an increase in fitness of the evolved populations and in the loss in the number of genes activated by quorum sensing. These results are consistent with the fact that many of the quorum sensing regulated traits are energetically costly.

They used a transcriptomics population approach, where they measured genes expression of the evolved population in the quorum sensing active and inactive conditions, and they used these results to quantify the number of genes activated in the evolved populations versus the ancestral. This is a novel approach that it is useful to study the effects of evolution at the population level.

The authors sequenced some of evolved clones and were able to obtain insights on the mechanist explanation for why one of the genes in the evolved population was no longer subject to quorum sensing regulation. By measuring promoter activity of a quorum sensing activated gene they could show that expression of this promoter was often decrease in some evolved clones while in others the promoter had delayed quorum sensing activation. These differences were shown to be related to different SNPs in the evolved clones in the regulator of this promoter. These results led to the conclusion that elimination of a gene from the quorum sensing can occur through different trajectories. I think it is interesting that reduction in the quorum sensing regulon can occur by small changes (SNPs) in the regulators belonging to the quorum sensing network because it might allow easy reversion of these losses. However, further characterization of other mutations from the evolved clones would be useful to understand how often these strategies occurred in opposition to others to understand the mechanistic explanation behind the loss of the genes from the quorum sensing regulon.

The authors wanted to obtain insight regarding why some strains have a much larger number of quorum sensing regulated genes than others, and proposed that the number of genes regulated by quorum sensing might provide clues about the evolutionary history of different strains from different environments. I think this is an interesting hypothesis, but further work is needed to gain support for such hypothesis. Here the authors observed that populations evolved under conditions where very few quorum sensing regulated traits are needed could increase their fitness by reducing the quorum sensing regulon. This result is interesting, but it is expected that unnecessary traits are counter selected when not needed. Therefore, these results open the door for future experiments to determine if the system is flexible enough to revert these losses, in other words it would be interesting to perform similar experiments with the evolved clones under conditions where more quorum sensing regulated traits are beneficial to see if reversion can be achieved. If reversion indeed occurs, it will be interesting to determine if such reversion would occur through recovery of quorum sensing regulation or other mechanisms.

Nonetheless, the current study provides novel and relevant insight on the characterization of quorum sensing genes activated by quorum sensing in Pseudomonas at the population level, and on the mechanisms involved during selection for reduction of the quorum sensing regulon.

Reviewer #2 (Public Review):

QS is a mechanism by which cells communicate with each other to coordinately regulate various behaviors. The question the authors attempt to address here is if these QS-regulated behaviors are no longer useful, will bacteria evolve to reduce the number of genes regulated by QS. Using the well-established P. aeruginosa model system, the authors show that after 1000 generations of evolution in a medium requiring QS for growth, the bacteria do reduce the number of genes induced by QS. They conclude this reduction increases the fitness of the evolved bacteria, although there is little data provided to support this claim. In addition, analysis of gene expression was carried out in a medium that is not reflective of the medium in which the evolution took place. This study is interesting as it combines long-term experimental evolution of a bacterial species in a medium that is dependent on maintaining QS with transcriptomic analysis of QS gene expression.

Reviewer #1 (Public Review):

The work by Ren et al. reveals a novel mechanism used by bone marrow-derived thymic seeding progenitors (TSPs) to engage thymic portal endothelial cells (TPECs) in order to gain access into the thymus. How TSPs enter the thymus is a longstanding question in the field. Here the authors show that SIRPa which is typically thought as a "don't eat me" receptor on macrophages, is expressed by TPECs and is engaged by CD47 expressed on TSPs to then signal TPECs to endocytose VE-cadherin, and thus facilitate transendothelial migration (TEM). The findings convincingly show that SIRPa is required by TPECs to enhance TSP entry into the thymus, and a role for SHP2 and Src signaling by SIRPa is supported using in an in vitro cell system to model TEM.

Reviewer #2 (Public Review):

The molecular mechanisms as well as the cellular players of colonization of the adult thymus are incompletely understood. In this manuscript, the authors investigate the role of the SIRPa-CD47 ligand pair in seeding of bone-marrow derived progenitors to the adult murine thymus. The study is based on the authors' earlier characterization of thymic portal endothelial cells, which have a role in mediating progenitor homing to the thymus (Shi et al., 2016). The authors show that loss of SIRPa or CD47 results in reduced frequencies and numbers of early T lineage progenitors (ETPs), but no substantial alterations in thymocyte numbers at later developmental stages and of bone-marrow precursors. Short-term homing assays suggest impaired colonization of the thymus. The authors further characterize cell biology and biochemistry of the SIRPa-CD47 system using peripheral lymphocyte co-cultures with genetically engineered MS1 endothelial cells. Finally, they assess the role of SIRPa-CD47 in thymus regeneration in combination with growth of a model tumor.

Strengths:

The authors describe a clear phenotype, consistent with the moderate effect size in ETP loss upon deletion of other homing mediators, such as PSGL-1 or individual chemokine receptors, such as CCR7, CCR9 or CXCR4.

The authors use multiple genetic models, including both, SIRPa and CD47 deficient mouse strains, to support their findings. Using the Tie2Cre model for endothelial cell-specific deletion is particularly informative and could have been used more extensively. Some data are further strengthened by the complementary use of inhibitory SIRPa-Ig fusion proteins.

In vitro analysis of the molecular mechanism and the role of signaling mediators using MS1 cells is well executed and conclusive.

Weaknesses:

Short-term homing assays suffer from the problem that the system is overwhelmed by an excessive number of donor cells (millions), whereas at steady state only a few hundred HPCs capable of colonizing the thymus circulate in peripheral blood, questioning the physiological relevance of this approach. The short-term nature of the experiments also precludes analysis, whether homed cells do in fact constitute T cell progenitors. More suitable experiments comprise mixed competitive bone marrow chimeras using congenically discernible donor cells or, even better, transfers into non-irradiated recipients of defined age as pioneered by the Goldschneider and Petrie labs. Thus, the conclusion that the SIRPa-CD47 system mediates homing of thymus seeding progenitors is not fully justified.

While technically elegant and mechanistically conclusive, the in vitro studies using MS1 cells and peripheral lymphocytes are somewhat isolated from the original focus of the paper addressing the role of SIRPa-CD47 specifically in thymus seeding. It should be considered devising similar assays replacing lymphocytes with bone-marrow derived progenitors.

Analysis of thymus regeneration is interesting, but a number of open questions remain for this experimental setup, also in part raised by the authors in the discussion section. Most notably, during regeneration, the reduction in ETPs is accompanied by reduced numbers in more mature thymocyte subsets and peripheral T cells. Such a reduction was not observed at steady-state in KO models and it cannot be concluded from this experiment, that these observations are caused by a defect in thymus colonization. Notably, SL-TBI is associated with massive cell death and alterations in phagocytosis and many other factors may come into play here as well.

Taken together, the study in its presents form contains the description of an interesting new phenotype, consistent with a role of the CD47-SIRPa interaction in colonization of the thymus by bone-marrow derived progenitors. However, at present, homing experiments lack sufficient rigor and experiments on thymus regeneration, while showing an interesting additional finding, do not justify to conclude homing as mechanistic explanation.

Reviewer #3 (Public Review):

The manuscript by Ren et al. seeks to describe a role for endothelial cell (EC) expression of Sirpα playing a role in the importation of hematopoietic progenitors from the circulation into the thymus. Specifically, the authors demonstrate that there is a reduction in the number of the earliest T lineage progenitors (ETPs) in the thymus in mice deficient for Sirpa or CD47 (its ligand), and through a series of elegant in vitro transendothelial migration studies, identify that intracellular Sirpα signaling mediates this process by regulating VE-Cadherin expression and thus EC tight junctions. In particular, the use of transwell assays modified to study TEM is particularly well utilized to tease apart the mechanisms. Overall, I found this to be an excellent manuscript. In fact, every time I had a critique developing in my head, the authors quickly dispensed of it by producing some follow up data that addressed my concern! My biggest concern with the manuscript is that it was difficult to determine exactly how many repeats of each experiment have been performed and what data is being presented in the figures (and being statistically analyzed). This should not change the conclusions of the manuscript but will make reading the figures and matching them with the legends easier. The following are a some major and minor concerns that should be addressed to strengthen the manuscript:

Major:<br /> • My main concern is that there needs to be greater care taken with highlighting the number of repeats done for each individual study as it is not always clear. For instance, in Figure 2 the data are presented as being representative of three independent experiments with an n of 3 in each experiment but in 2B, D, and F there are 4 data points for the Sirpa-/- group. This is likely explained by there being 4 mice in that particular experiment, but that is why the numbers should be presented for each experiment rather than a general statement at the end. Another example of this is that in Figure 2 S1 the authors would like to claim that the only differences are in the DN1 subsets which contains the ETPs. However, it is likely this is just due to low numbers as it seems like there is a real decrease in the number of DN2, DN3, DN4 and even DP thymocytes (as well as total cellularity).<br /> 1. This should not change any conclusions of the paper but will aid in reader interpretation.<br /> 2. In this manuscript the authors show that Sirpa expression by TPECs is critical for their capacity to guide the importation of HPCs, and in their previous work they have shown that lymphotoxin can regulate the importation capacity of these same TPECs. Therefore, it would be extremely interesting to know if LT signaling is regulating the expression of Sirpa. Furthermore, it would be important to at least comment on what may be influencing Sirpa expression. For instance, we know from the work of Petrie and others that DN niche availability can influence the ability of the thymus to import of progenitors. Similarly, after TBI the "gates" are let open and the capacity of the thymus to import progenitors increases. Do the authors know (or could they comment) on what happens to Sipra expression after TBI in ECs?<br /> 3. The use of the in vitro TEM assays in transwell plates are a nifty way of interrogating and manipulating the effect of Sirpa in these conditions, however, the caveat is that these all use EC cell lines that do not correspond to the TPECs being described in vivo. This caveat should be acknowledged in the text.<br /> 4. I am a little confused as to the interpretation of the final experiment looking at tumor clearance. The authors show that this could be clinically relevant as blockade of the CD47-Sirpa axis is becoming an increasingly attractive immunotherapy option but its use could preclude thymic recovery after damage and thus contribute toward poorer T cell responses against tumors. This last study is very interesting but also very hard to interpret given the likely positive effect of Sirpa-CD47 blockade on tumor clearance, in opposition to its potential effects hindering thymic repair. While it is notable that there is reduced clearance of tumor in mice treated with CV1, it is unclear why there does not seem to be any positive effect of CV1 on tumor clearance (is this because there are fewer T cells in the periphery as it is still early after damage?). On the thymic repair and reconstitution front, perhaps a cleaner way would be to look in Sirpa or CD47 deficient mice and without tumors.

Minor Comments:<br /> • In Fig. 2I (and Fig. 2S2I-J), it is difficult to determine how long after the chimera transplant the homing assays were performed. However, this approach has limitations as the process of creating those chimeras (conditioning such as irradiation etc.) will change the function and possibly the mechanisms of progenitor entry into the thymus. There is clearly still an effect of Sirpa in this context but it is possible (even likely) that the importation mechanisms in the thymus change after damage such as that caused by the conditioning required in the initial chimera generation. Furthermore, although using the Tie2-Cre strain will distinguish Sirpa on ECs and TECs, it will not distinguish between expression on other cells such as DCs (Tie2 will delete expression in both endothelial and hematopoietic lineages). Although the optimal experiment to address these concerns would be to delete Sirpa from ECs specifically (such as with Cdh5-CreERT2 mice), I am convinced by the preponderance of in vitro data that there is an EC-specific effect and therefore it is not necessary to perform this time-consuming, albeit interesting, potential experiment. However, these limitations should be acknowledged in the discussion or text.<br /> • As a technical note I am surprised that there was considerable reconstitution of naive T cells at day 21 after TBI (Fig.7G-H). In our experience that is very early for naïve T cells in the periphery which generally take about 4 weeks to start reconstituting in a real sense. Is it possible there are direct effects of this treatment on residual radio-resistant peripheral T cell numbers?

Reviewer #1 (Public Review):

Beier leveraged a more selective Rabies virus retrograde tracing method to identify, for the first time, local monosynaptic inputs to DAT+ dopamine neurons in the VTA. He identified a previously unappreciated preponderance of SNr inhibitory inputs, as well as local connectivity among neighboring DA neurons and inputs from serotonergic inputs of the raphe nucleus. Rigorous neuroanatomy is a prerequisite to identify how DA neurons compute reward, prediction, and movement signals during behavior.

Reviewer #2 (Public Review):

This single-author study highlights an important caveat of currently prevalent monosynaptically-restricted rabies tracing methods in neuroscience, which complicate the interpretation of data on synaptic inputs surrounding injection sites. The author used a previously published, modified viral strategy (with weakened TVA expression) to overcome this shortcoming and revisit both global and local inputs to dopaminergic neurons of the ventral tegmental area. This study provides a useful tool for the field, along with significant new data, essentially correcting/updating a series of datasets from high profile papers originally mapping the inputs to DA neurons. This work can be strengthened in a few ways to enhance clarity and impact and can improve in addressing alternative, previous methods that may solve similar problems.

Related to Figure 1:<br /> Previously published studies have attempted to address the problem of low background levels of TVA expression and non-Cre dependent labeling of inputs near injection sites. One solution to prevent this ectopic expression of TVA is to titrate concentration of helper virus for efficient labeling of starter cells while minimizing non-Cre mediated expression, e.g. Wickersham's group https://doi.org/10.3389/fnsyn.2020.00006. Overall, there is an omission of discussion of alternative, potentially far simpler solutions. For example, how do volume, titer, AAV serotype, promoter, and other batch effects could alter relative to AAV helper virus expression change the background level of non-Cre dependent TVA infection? Without varying these parameters extensively, especially given the diverse applications of this technology in hundreds of neuroscience labs, it is challenging to evaluate the depth of the problem with mapping local inputs.

Related to Figure 2:<br /> The author's use of dimensionality reduction to analyze the relationship across local and long-range inputs is compelling and offers new ways of dissecting large-scale monosynaptic rabies tracing data. The author used z-score of fractional counts for UMAP input data and this ultimately aligns to spatial location of input fibers within the VTA. Considering that the UMAP input data is fractional counts for the number of input cells, does the corresponding spatial location of fibers relate to the density of fibers or fibers that are contacting specific subsets of VTA neurons?

Related to Figure 3:<br /> The author utilizes his altered viral strategy to dissect local inhibition to VTA dopamine neurons. This method provides critical new quantitative information about sources of local inhibition within the VTA, but the information would be significantly strengthened in its impact if complemented by immunohistochemistry for specific cell type markers of GABAergic cell types. For example, the use of other probes beyond GAD with UMAP analyses might begin to answer whether subsets of DA neurons receive different quantity of inhibitory synaptic contacts or receive different types of inhibition from specific inhibitory cell types.

Related to Figure 4:<br /> Why is immunofluorescence used to study DA-DA connectivity as opposed to FISH as used for GABAergic input? They are essentially the same type of analyses but two different methodologies are used without justification. Why was this analysis restricted to only VTA (or neighboring) regions when other long-range Th+ DA neurons, such as those in the hypothalamus, might also make connections to VTA DA neurons? While the focus on local inputs makes sense in the context of the study and TVA spread, there is a missed opportunity to delve deeper and create a unique resource that contributes biological insight on TH+ neuronal network organization.

Reviewer #3 (Public Review):

A number of prior studies have used variants of cell-type specific rabies viral tracing to examine the inputs to midbrain dopamine neurons. This method generally speaking requires cells to express both the TVA protein that allows transsynaptic viral jumps as well as infection with and expression of a rabies viral backbone expressing a reporter. Although a transsynaptic jump is relatively rare (estimated here and elsewhere to be ~10 presynaptic jumps per double-infected neuron), the very modest leak of in cell-type specific expression using the cre recombinase system can lead to small amounts of TVA expression in nominally non-cre expressing cells (in this case non-dopaminergic neurons) and some transsynaptic labeling as a result. In this study Beier takes advantage of a recently described point mutant of TVA that reduces efficiency of transsynaptic labeling and thus requires higher levels of TVA expression for effect labelling. As a result nominally non-cre expressing cells are even less likely to mediate transsynaptic labeling. Although the number of experiments is relatively small (3-4) there is quite a clear reduction in off-target infection using this strategy (very similar to background levels observed in mice with no cre expression). Beier identifies a particularly important use case for this approach - namely assessing inputs to dopaminergic neurons from non-dopaminergic neurons within the area of viral injection. Due to the relatively large fraction of non-dopaminergic neurons in the midbrain previous uses of rabies transsynaptic tracing methods can be overwhelmed by off-target expression. This is particularly relevant because it has long been known from detailed electron microscopy analysis and functional studies of synaptic input that dopamine neurons have substantial, predominant inhibitory inputs. However, inhibitory inputs are, in general, biased towards more anatomically local sources in general in the brain and also in the midbrain. Thus, prior work using similar tracing methods was biased towards long range excitatory inputs to midbrain dopamine neurons. Beier here provides an important, valuable, and complementary tracing dataset that likely more accurately captures the input to dopamine neurons locally and yet is also consistent with prior long range tracing data. The manuscript is relatively focused on comparing this work with prior anatomical work, especially that using rabies mediated transsynaptic tracing. This is understandable given its focus, however, a number of functional studies and other anatomical studies have made similar important points about the relevance of local connectivity and yet those studies were little discussed. The dataset is also derived from a small set of injections (4) and focused on a particular coordinate in lateral VTA with some labelling in medial SNc and more medial VTA. Thus some questions remain about how well these results are representative of the broader population of midbrain dopamine neurons and more subtle differences in local connectivity. The "convergence index" is similar here to studies with TVA (~7-8) presumably due to improved RABV efficiency that compensates for a partial LOF mutation in TVA. Nonetheless, the number of input cells labelled per postsynaptic dopamine neuron is still a very small fraction of total inputs raising some questions about the weather heterogeneity of input patterns is accurately assessed. Finally, potential tropism for specific input cell classes remains an unknown.

Reviewer #2 (Public Review):

Overview:

This paper explores the use of inflow contrast MRI for imaging the pial arteries. The paper begins by providing a thorough background description of pial arteries, including past studies investigating the velocity and diameter. Following this, the authors consider this information to optimize the contrast between pial arteries and background tissue. This analysis reveals spatial resolution to be a strong factor influencing the contrast of the pial arteries. Finally, experiments are performed on a 7T MRI to investigate: the effect of spatial resolution by acquiring images at multiple resolutions, demonstrate the feasibility of acquiring ultrahigh resolution 3D TOF, the effect of displacement artifacts, and the prospect of using T2* to remove venous voxels.

Impression:

There is certainly interest in tools to improve our understanding of the architecture of the small vessels of the brain and this work does address this. The background description of the pial arteries is very complete and the manuscript is very well prepared. The images are also extremely impressive, likely benefiting from motion correction, 7T, and a very long scan time. The authors also commit to open science and provide the data in an open platform. Given this, I do feel the manuscript to be of value to the community; however, there are concerns with the methods for optimization, the qualitative nature of the experiments, and conclusions drawn from some of the experiments.

Specific Comments:

1. Figure 3 and Theory surrounding. The optimization shown in Figure 3 is based fixing the flip angle or the TR. As is well described in the literature, there is a strong interdependency of flip angle and TR. This is all well described in literature dating back to the early 90s. While I think it reasonable to consider these effects in optimization, the language needs to include this interdependency or simply reference past work and specify how the flip angle was chosen. The human experiments do not include any investigation of flip angle or TR optimization.

2. Figure 4 and Theory surrounding. A major limitation of this analysis is the lack of inclusion of noise in the analysis. I believe the results to be obvious that the FRE will be modulated by partial volume effects, here described quadratically by assuming the vessel to pass through the voxel. This would substantially modify the analysis, with a shift towards higher voxel volumes (scan time being equal). The authors suggest the FRE to be the dominant factor effecting segmentation; however, segmentation is limited by noise as much as contrast.

3. Page 11, Line 225. "only a fraction of the blood is replaced" I think the language should be reworded. There are certainly water molecules in blood which have experience more excitation B1 pulses due to the parabolic flow upstream and the temporal variation in flow. There is magnetization diffusion which reduces the discrepancy; however, it seems pertinent to just say the authors assume the signal is represented by the average arrival time. This analysis is never verified and is only approximate anyways. The "blood dwell time" is also an average since voxels near the wall will travel more slowly. Overall, I recommend reducing the conjecture in this section.

4. Page 13, Line 260. "two-compartment modelling" I think this section is better labeled "Extension to consider partial volume effects" The compartments are not interacting in any sense in this work.

5. Page 14, Line 284. "In practice, a reduction in slab ...." "reducing the voxel size is a much more promising avenue" There is a fair amount on conjecture here which is not supported by experiments. While this may be true, the authors also use a classical approach with quite thin slabs.

6. Figure 5. These image differences are highly exaggerated by the lack of zero filling (or any interpolation) and the fact that the wildly different. The interpolation should be addressed, and the scan time discrepancy listed as a limitation.

7. Figure 7. Given the limited nature of experiment may it not also be possible the subject moved more, had differing brain blood flow, etc. Were these lengthy scans acquired in the same session? Many of these differences could be attributed to other differences than the small difference in spatial resolution.

8. Page 22, Line 395. Would the analysis be any different with an absolute difference? The FRE (Eq 6) divides by a constant value. Clearly there is value in the difference as other subtractive inflow imaging would have infinite FRE (not considering noise as the authors do).

9. Page 22, Line 400. "The appropriateness of " This also ignores noise. The absolute enhancement is the inherent magnetization available. The results in Figure 5, 6, 7 don't readily support a ratio over and absolute difference accounting for partial volume effects.

10. Page 24, Line 453. "strategies, such as radial and spiral acquisitions, experience no vessel displacement artefact" These do observe flow related distortions as well, just not typically called displacement.

11. Page 24, Line 272. "although even with this nearly ideal subject behaviour approximately 1 in 4 scans still had to be discarded and repeated" This is certainly a potential source of bias in the comparisons.

12. Page 25, Line 489. "then need to include the effects of various analog and digital filters" While the analysis may benefit from some of this, most is not at all required for analysis based on optimization of the imaging parameters.

Reviewer #1 (Public Review):

In this article, Bollmann and colleagues demonstrated both theoretically and experimentally that blood vessels could be targeted at the mesoscopic scale with time-of-flight magnetic resonance imaging (TOF-MRI). With a mathematical model that includes partial voluming effects explicitly, they outline how small voxels reduce the dependency of blood dwell time, a key parameter of the TOF sequence, on blood velocity. Through several experiments on three human subjects, they show that increasing resolution improves contrast and evaluate additional issues such as vessel displacement artifacts and the separation of veins and arteries.

The overall presentation of the main finding, that small voxels are beneficial for mesoscopic pial vessels, is clear and well discussed, although difficult to grasp fully without a good prior understanding of the underlying TOF-MRI sequence principles. Results are convincing, and some of the data both raw and processed have been provided publicly. Visual inspection and comparisons of different scans are provided, although no quantification or statistical comparison of the results are included.

Potential applications of the study are varied, from modeling more precisely functional MRI signals to assessing the health of small vessels. Overall, this article reopens a window on studying the vasculature of the human brain in great detail, for which studies have been surprisingly limited until recently.

In summary, this article provides a clear demonstration that small pial vessels can indeed be imaged successfully with extremely high voxel resolution. There are however several concerns with the current manuscript, hopefully addressable within the study.

Main points:

1. The manuscript needs clarifying through some additional background information for a readership wider than expert MR physicists. The TOF-MRA sequence and its underlying principles should be introduced first thing, even before discussing vascular anatomy, as it is the key to understanding what aspects of blood physiology and MRI parameters matter here. MR physics shorthand terms should be avoided or defined, as 'spins' or 'relaxation' are not obvious to everybody. The relationship between delivery time and slab thickness should be made clear as well.

2., The main discussion of higher resolution leading to improvements rather than loss presented here seems a bit one-sided: for a more objective understanding of the differences it would be worth to explicitly derive the 'classical' treatment and show how it leads to different conclusions than the present one. In particular, the link made in the discussion between using relative magnetization and modeling partial voluming seems unclear, as both are unrelated. One could also argue that in theory higher resolution imaging is always better, but of course there are practical considerations in play: SNR, dynamics of the measured effect vs speed of acquisition, motion, etc. These issues are not really integrated into the model, even though they provide strong constraints on what can be done. It would be good to at least discuss the constraints that 140 or 160 microns resolution imposes on what is achievable at present.

3. The article seems to imply that TOF-MRA is the only adequate technique to image brain vasculature, while T2* mapping, UHF T1 mapping (see e.g. Choi et al., https://doi.org/10.1016/j.neuroimage.2020.117259) phase (e.g. Fan et al., doi:10.1038/jcbfm.2014.187), QSM (see e.g. Huck et al., https://doi.org/10.1007/s00429-019-01919-4), or a combination (Bernier et al., https://doi.org/10.1002/hbm.24337​, Ward et al., https://doi.org/10.1016/j.neuroimage.2017.10.049) all depict some level of vascular detail. It would be worth quickly reviewing the different effects of blood on MRI contrast and how those have been used in different approaches to measure vasculature. This would in particular help clarify the experiment combining TOF with T2* mapping used to separate arteries from veins (more on this question below).

4. The results, while very impressive, are mostly qualitative. This seems a missed opportunity to strengthen the points of the paper: given the segmentations already made, the amount/density of detected vessels could be compared across scans for the data of Fig. 5 and 7. The minimum distance between vessels could be measured in Fig. 8 to show a 2D distribution and/or a spatial map of the displacement. The number of vessels labeled as veins instead of arteries in Fig. 9 could be given. In the main quantification given, the estimation of FRE increase with resolution, it would make more sense to perform the segmentation independently for each scan and estimate the corresponding FRE: using the mask from the highest resolution scan only biases the results. It is unclear also if the background tissue measurement one voxel outside took partial voluming into account (by leaving a one voxel free interface between vessel and background). In this analysis, it would also be interesting to estimate SNR, so you can compare SNR and FRE across resolutions, also helpful for the discussion on SNR.

5. The separation of arterial and venous components is a bit puzzling, partly because the methodology used is not fully explained, but also partly because the reasons invoked (flow artefact in large pial veins) do not match the results (many small vessels are included as veins). This question of separating both types of vessels is quite important for applications, so the whole procedure should be explained in detail. The use of short T2* seemed also sub-optimal, as both arteries and veins result in shorter T2* compared to most brain tissues: wouldn't a susceptibility-based measure (SWI or better QSM) provide a better separation? Finally, since the T2* map and the regular TOF map are at different resolutions, masking out the vessels labeled as veins will likely result in the smaller veins being left out.

6. A more general question also is why this imaging method is limited to pial vessels: at 140 microns, the larger intra-cortical vessels should be appearing (group 6 in Duvernoy, 1981: diameters between 50 and 240 microns). Are there other reasons these vessels are not detected? Similarly, it seems there is no arterial vasculature detected in the white matter here: it is due to the rather superior location of the imaging slab, or a limitation of the method? Likewise, all three results focus on a rather homogeneous region of cerebral cortex, in terms of vascularisation. It would be interesting for applications to demonstrate the capabilities of the method in more complex regions, e.g. the densely vascularised cerebellum, or more heterogeneous regions like the midbrain. Finally, it is notable that all three subjects appear to have rather different densities of vessels, from sparse (participant II) to dense (participant I), with some inhomogeneities in density (frontal region in participant III) and inconsistencies in detection (sinuses absent in participant II). All these points should be discussed.

7. One of the main practical limitations of the proposed method is the use of a very small imaging slab. It is mentioned in the discussion that thicker slabs are not only possible, but beneficial both in terms of SNR and acceleration possibilities. What are the limitations that prevented their use in the present study? With the current approach, what would be the estimated time needed to acquire the vascular map of an entire brain? It would also be good to indicate whether specific processing was needed to stitch together the multiple slab images in Fig. 6-9, S2.

8. Some researchers and clinicians will argue that you can attain best results with anisotropic voxels, combining higher SNR and higher resolution. It would be good to briefly mention why isotropic voxels are preferred here, and whether anisotropic voxels would make sense at all in this context.

Reviewer #1 (Public Review):

In this manuscript, the authors exploit retinal cell proliferation and neurogenesis in zebrafish to study banp, a protein that is essential in humans and embryonic lethal in mice. The authors performed large-scale mutagenesis and identified a mutant known as "rw337" that compared to WT cells the mutant zebrafish have smaller eyes and optic tectum. They found that the retinas of these mutants have mitotic-like round cells that accumulate indicating mitotic arrest. Sequencing of these mutants identified that the rw337 mutant gene encodes a truncated banp protein. Expression of WT Banp occurs primarily in retinal and neuronal cells in Zebrafish. Interestingly, rw337 showed significant decrease in retinal photoreceptors number and neuronal formation within the OPL and IPL were morphologically disrupted and had fewer cells. The authors found that rw337 cells have increased numbers of DSBs in the retina over time (via TUNEL) assays. They found that mitotic defects and apoptosis are spatially and temporally occurring in distinct regions of the retina as prolonged phosphorylation of histone H3, which indicates an issue in exit of mitosis, occurred in apical surface of the neural retina whereas apoptosis occurred in retinal progenitor cells (via Caspase 3 staining). The authors then went on to examine the role of replication stress regulators like p53, atm, and atr and showed that protein and RNA levels of banprw337 were increased and upregulated. As p53 binds banp in zebrafish, it was not surprising that regulators of p53 were enhanced in banprw337 mutants. Intriguingly, the authors found that two genes which are essential for chromatin segregation were downregulated in banprw337 mutants and banp morphants as a result of chromatin accessability decreases near the TSS of resulting in decreased transcriptional activity of cenpt and ncapg genes. Finally, the authors temporally monitored mitosis in mitosis of banprw337 mutants and found that chromosomal segregation is abnormal and takes longer. The authors have performed a thorough analysis of the impact of the banp gene on retinal biology and its importance regulating replication stress response and cenpt and ncapg expression. This paper is important to retinal biology, genome stability, and replication stress response fields and requires minor revision.

Strengths:<br /> • These studies exploit zebrafish retinal development and its cell-cycle regulation as knockout of Banp/ SMAR1 is an essential gene in human cells and embryonic lethal in mice.<br /> • The authors show that this gene is involved in replication stress responses involving p53, atm, and atr signaling.<br /> • The authors show that banp is required for chromatin segregation factors and chromatin accessability by binding to banp sequences (TCTCGCGAGA) upstream of specifically cenpt and ncapg. Interestigly the mutant rw337 had decreased chromatin accessability near the transcript start sites of these genes. This is an elegant study of how a gene is regulating the transcription of two genes essential for chromatin segregation.<br /> •<br /> Weaknesses:<br /> • The authors could highlight the protein names of both zebrafish and humans throughout the text using standard nomenclature description with humans proteins all capitalized etc... This will enable the reader to understand their findings in the context of fascinating biology and human disease/cancer.<br /> • As banprw337 mutants show such severe morphological disruption a discussion on the impact of this work for the vision community could strengthen the importance of understanding how this gene functions.<br /> • Gamma H2AX phosphorylation is a global marker of DSBs and stalled forks. The authors did not note that H2AX phorylation is present and a marker of stalled replications forks.<br /> o PMID: 11673449, PMID: 20053681, doi:10.1101/gad.2053211, https://doi.org/10.1016/j.cell.2013.10.043 etc.<br /> • As gamma H2AX phosphorylation recruits DNA repair factors like BRCA2, speculation of importance of these genes may be of interest to the DNA repair community.

Reviewer #2 (Public Review):

Babu et al report the role of the zebrafish banp gene in the developing retina. They find that banp is required for faithful S-phase as well as mitosis.

Manuscript strengths:<br /> 1- The authors performed a large-scale mutagenesis screen and successfully identified a causative banp gene mutation from these efforts, which represent a significant amount of work.<br /> 2- The authors provide a substantial amount of cellular-level analysis of a host of cell cycle-related phenotypes in the banp mutant retina. The data are of high technical quality and the experiments are well-executed. For the most part, the data support the conclusions.

Manuscript weaknesses:<br /> 1- Banp mutants have numerous defects, and perhaps this is not unexpected for a nuclear matrix protein. I'm left wondering what insights are gained from the study beyond that the nuclear matrix is required for numerous cell cycle events?<br /> 2- Why did the authors focus on the eye? It is unclear whether this study revealed a sensitivity to eye development regarding nuclear matrix function specifically, or it was just a convenient place in the animal to look.<br /> 3- I found the conclusions regarding mitosis to be contradictory. The authors at first emphasize mitotic arrest, but then characterize chromosome segregation defects. How can chromosomes segregate if cells are arrested in mitosis?<br /> 4- It would be important to know whether the authors can rule out that S-phase defects cause the M phase defects, or vice versa. Could there be a primary defect, rather than multiple independent defects as the authors conclude?

Reviewer #3 (Public Review):

Babu and colleagues demonstrate that banp is expressed in the retina progenitor cells among other locations, and mutational loss of it results in increased mitosis, increased apoptosis, increased DNA damage, and the failure to differentiate photoreceptors. Importantly, these phenotypes are seen at a time period when retina progenitors undergo rapid cell cycles and differentiate into multiple cell types that make up the fully developed retina. Rescue with the wild type and phenocopy with another mutant allele provide strong support that the phenotypes results from loss of banp. Mutant animals show elevated p53 protein and reduction of p53 delays the onset of apoptosis by 24 hours. Mutant animals show altered transcriptional profile, with increased p53 expression and decreased expression of two genes that encode proteins needed for chromosome segregation. The authors propose that loss of banp results in defective DNA replication and DNA damage as well as mitotic chromosome segregation failures, all of which contribute to p53-dependent apoptosis to reduce cell number and cause developmental defects.

Banp is a very interesting protein. Also known as Scaffold/matrix attachment region binding protein 1, it is known to regulate the transcription of a number of genes including those important in oncogenesis. In vivo function of Banp, especially in the context of normal development, remains to be better understood. The current study fills this knowledge gap but I have some concerns about the interpretation of the data, the presentation and the potential impact. Specifically:

Increased expression of atm and atr is observed and the authors suggest that replication stress and DNA damage activate the checkpoints to cause cell cycle arrest. There are several problems with this conclusion, which is depicted in Fig. 4G. Checkpoint activation occurs via phosphorylation changes in ATM/ATR and not through their transcriptional upregulation, which would take too long for a response that occurs within minutes. ATM/ATR-dependent checkpoints arrest cells in G1 or G2 so you would expect reduced S and M phases. Yet, the authors saw increased M and no change in S.

It is puzzling that BrdU+ cell number does not change because if cells are indeed arrested in mitosis, they should be prevented from going into S phase and BrdU+ cell numbers should decrease.

It is not addressed whether cenpt and ncapg expressed in the retina and whether are their expressions decreased in banp mutants. The RNAseq data is from whole animals.

The rescue by banp-EGFP in Fig.1G is very nice. But it looks like there is partial rescue also with EGFP-banp(rw337) in the same panel. The defects the last panel do not seem as severe as in non inj controls. There are fewer pyknotic nuclei and the cell layers lack gaps. Quantification of the extent or reproducibility of the rescue is lacking.

Some of the conclusions lack supporting data. For example, line 99: "Thus, Banp is required for integrity of DNA replication and DNA damage repair." There are no data for the integrity (meaning 'fidelity'?) of DNA replication and there are no DNA repair assays. In another example, non-overlap of pH3 (M phase) and caspase+ cells is interpreted to mean that cells are dying in S phase (Figure 2 supplement 1). But the data are equally consistent with cells dying in G1 and G2.

The model in Figure 7 includes components without accompanying supportive data. For example, the arrow from Banp to DNA repair that indicates a direct role and the arrow from tp53 to delta113 tp53 that indicates direct activation.

The data that together support a single point are often split up among figures. For example, increased pH3+ cells shown in Fig. 2 and is interpreted as mitotic arrest. But it is equally possible that cells are undergoing extra divisions (and then dying). Support for mitotic arrest is provided by live imaging of mitosis, which is not presented until the last figure (Fig. 6). There are many such instances in the manuscript.

Banp is already known for roles in p53-dependent transcription and in apoptosis (e.g. Sinha et al papers cited in the manuscript). Banp is also known to bind to the promoter regions of cenpt and ncapg (Grand et al and Mathai et al papers cited in the manuscript). These genes are known to be involved in mitosis in zebrafish (Hung et al and Seipold et al papers cited in the manuscript). In terms of what is new about banp function in this report, the requirement for banp in a critical phase of retina development and spontaneous induction of DNA damage come to mind. Unfortunately, how loss of banp leads to this defect remains to be addressed.

Reviewer #1 (Public Review):

Zhao and colleagues investigated the multimodal signalling system of the small torrent frog (Amolops torrentis) and its potential link to ecto-parasites (e.g. blood-sucking midges). The authors present an innovative taxon-oriented behavioural field study. I do not know any similar study and I welcome the authors' effort to describe the multimodal signals of A. torrentis in such great detail. They find that male small torrent frogs display a rich repertoire of visual cues, such as wiggling fingers or toes, or stretching and flagging front or hind legs, in addition to movement and sounds associated with advertisement calls. They show that males display more conspicuous visual displays, when conspecific females are in close proximity, and that those displays do influence the behaviour of the female receivers.

Another intriguing result of the present study is the demonstrated origin of the limb displays. Zhao et al. find that the original function of the diverse limb movements is to fend off blood-sucking parasites. However, these exact movements have been shown to increase the attractiveness of a male's display to female receivers when combined with the advertisement call: a novel result in the field of anuran bio-acoustics, and especially interesting, as the study species is endemic to only two regions in China and currently threatened by habitat loss.

The authors conducted an impressive study combining observations, audio and video recordings and behavioural experiments in order to reveal the signalling strategies of A. torrentis. The manuscript provides a fascinating example of a potentially highly complex signalling system in an anuran amphibian -- still a vastly understudied area in the field of animal communication. The study design is well thought-out and I cannot detect any faults in the analysis. The authors interpret their data modestly and draw reasonable conclusions. The figures are well presented and self-explanatory. The scientific illustrations and videos help to understand the fascinating signalling behavior. I really enjoyed reading this manuscript and I am sure, many colleagues will too.

Reviewer #2 (Public Review):

Organisms communicate using different sets of signals, ranging from visual, acoustic, chemical, to tactile cues. Such signals transfer information about the signaler to an intended receiver that decodes such cues. Natural selection (and sexual selection) usually can shape how exclusive such signals become for their intended receivers. Depending on the context, some signals might include more than one information channel and become multimodal (e.g., visual + acoustic). Such complex signals might allow receivers to better decode information about the signaler (e.g., genetic quality, health, resources). However, most signalers avoid being detected by predators in their habitat, thus there is generally a conflict between signal detectability by partners and predation avoidance by signalers. In anurans (frogs and toads), the dominant mating signals are acoustic, but visual cues can be combined in the courtship display if the signalers are active during the day and combine those with their vocalizations.

In the present study, the authors focused on the evolution of mating signals in torrent frogs (Amolops torrentis) from noisy mountain streams of the Hainan Island (China). Males of this species are active during day and night; and their displays include acoustic and visual cues that attract females. The authors show that the acoustic signals might also attract unintended parasites; in this case, midges (small flies) that locate vocalizing males and try to get a blood-meal. Such interaction seems to incite the males to move their limbs as if they are trying to swat away the parasites. By studying such displays, the authors show that body movements in males might increase female preference for them in addition of the acoustic signal. The authors hypothesize that limb movement has been adopted as part of a multimodal mating display where limb movements and vocalizations further entice female preference. For this purpose, the researchers filmed frogs in the field and classified their limb displays. Then, the authors showed evidence that, when males call, these animals tend to attract more midge parasites and such interaction increased the frequency of limb movements as males try to swat away these midges. Interestingly, such body movements, including foot-flagging and leg-stretching, if combined with mating calls, seem to make males more attractive to females.

This study provides an intriguing hypothesis -- namely, that mating displays might become more attractive if males are engaged in antiparasitic movements. These alleged visual cues capture the attention of females and effectively enhance the sensory components of this mating display. However, it is not clear how much of this antiparasitic (swatting) behavior has become a component of the mating display. For instance, such body movements might be a rather chaotic visual display in a vocalizing male that is trying to "scare away" the parasitic flies. Most amphibians have mating displays that are usually structured (i.e., a specific sequence of components in the signal) and partners can decode such sequences of signal components, which are genetically determined (i.e., they are not learned). I consider this study a nice natural history report, yet it is not conclusive on the integration of acoustic vocalization and defensive body movements as parts of a multimodal mating display. In other words, such body movements are not coevolving with acoustic signals, rather they are capturing the attention of females as a by-product of males that are frantically trying to scare away flies while vocalizing. I believe that vocalizations are the main signal that females are paying the most attention to, to evaluate the quality of potential male partners.

Reviewer #1 (Public Review):

In this manuscript, Guo et al. seek to understand how DSB formation is regulated during C. elegans meiosis. Building upon previous findings (including the lab's own previous published results), the authors focused on the balance between kinase and phosphatase activity on DSB-1, a key DSB protein and orthologue of the SPO-11 co-factor Rec114

Using a combination of genetic approaches, imaging and phospho-induced electrophoretic mobility shifts, the authors show that phosphorylation of DSB-1 is mediated by two major kinases, ATR and ATM, and is counteracted by the PP4 phosphatase, PPH-4.1. While PPH-4.1 was known to be involved in pairing, synapsis and generation of DSB, the authors report that an atl-1 (ATR) null (partially) rescues the DSB defect in pph-4.1 mutants. In line with this, a DSB-1 phospho-mutant lacking five S/T-Q sites rescues the homologous pairing and synapsis defect of pph-4.1 mutants. These results argue in favour of the notion that DSBs strengthen synapsis in C. elegans.

Reviewer #2 (Public Review):

In this paper, Guo et al. investigated how DNA double-strand break (DSB) formation is regulated during C. elegans meiosis. Meiotic recombination initiates with programmed DSB formation, which is catalyzed by the Spo11 holoenzyme. In C. elegans, three SPO-11 cofactors have been identified so far. One of them is DSB-1, which is one of two homologs of Rec114. The authors first show that a phosphorylated form of DSB-1 appears as a slower migrating species on western blots. Using this as a readout, they demonstrate that phosphorylation of DSB-1 is dependent on two DNA damage sensor kinases, ATR (ATL-1) and ATM (ATM-1) and that dephosphorylation of DSB-1 is partially mediated by a member of PP4 phosphatase, PPH-4.1. It was previously shown that PPH-4.1 is required for multiple steps in meiotic chromosome dynamics, such as homolog pairing, synapsis, and DSB formation. Interestingly, heterozygous null mutation of atl-1, but not atm-1 deletion, restores meiotic DSB formation in pph-4.1 animals. It was further shown that DSB-1 contains five S/T-Q sites within its disordered region, and mutating all five sites leads to increased DSB formation and partially restores homologous pairing, DSB formation, and chiasma formation in pph-4.1 mutants. The rescue of homolog pairing was unexpected, and this illustrates a requirement of meiotic DSBs in enforcing correct pairing in C. elegans, similar to the case in other eukaryotes. The authors further demonstrate that DSB-1 phosphorylation occurs in an age-dependent manner, and this trend is not observed in dsb-2 mutants, which leads to a proposal that DSB-2 might have evolved to compensate for the decreased activity of the phosphorylated DSB-1 in older animals.

Overall, this is a nice study illustrating the antagonistic relationship between ATL-1and PPH-4.1 in regulating meiotic DSB formation. This work establishes that meiotic DSB formation is negatively regulated by ATL-1 in C. elegans, similar to what has been established in other organisms, and adds that a PP4 family member opposes this function of ATR. Rescued DSB formation and homolog pairing in pph-4.1; dsb-1(5A) is striking (even though it's partial), indicating that DSB-1 is a major target of PPH-4.1. Perhaps the partial rescue is somewhat expected, as PPH-4.1 is involved in many meiotic processes other than DSB formation. Therefore, more thorough analyses of the "rescued" phenotypes in pph-4.1 mutants, especially the status of SC assembly in both pph-4.1; dsb-1(phosphomutant series) and irradiated animals (with different doses), will help clarify some of the discussion points regarding the function of PPH-4.1 in processing recombination intermediates and the degree to which exogenous DSBs contribute to homolog pairing and synapsis in C. elegans. Another criticism is that this study only focuses on the phosphoregulation of DSB-1, while both DSB-1 and DSB-2 are C. elegans homologs of Rec114, and DSB-2 also contains four S/T-Q sites. Structural prediction of the putative DSB-1:DSB-2:DSB-3 and DSB-1:DSB-1:DSB-3 complexes in the discussion is very illuminating and suggests that perhaps the remaining DSB-1 in dsb-2 mutants is the pool that forms the DSB-1:DSB-1:DSB-3 complex and is prematurely phosphorylated by ATL-1 simply because of mass action. A model figure illustrating the phospho-regulation of DSB-1 (and maybe DSB-2) by ATL-1 and PPH-4.1 will greatly strengthen this paper.

Reviewer #3 (Public Review):

In this paper, Guo et al describe a novel function that modulates the frequency of meiotic double-strand breaks that initiate recombination through the inhibition by phosphorylation of one protein necessary for DSB formation, DSB-1.<br /> Studies in several species (mammals, budding yeast, Drosophila) have described mechanisms that specifically limit the number of meiotic DSBs.

This study performed in C. elegans reports convincingly that there is a balance between phosphorylation of DSB-1 (mainly by the ATL-1 (ATR) kinase) and its dephosphorylation by the PPH-4.1 (PP4) conserved phosphatase. Furthermore, they propose a mechanism by which DSB-1 phosphorylation is evolving with age, and how it is balanced by the activity of a related DSB protein, DSB-2. Finally, using structure prediction, they propose a mechanism for the relative activities of the related DSB-1 and DSB-2 proteins within the protein complex required for DSB formation.

All the involved proteins are conserved, and interestingly, the mechanisms described here are reminiscent of results obtained in budding yeast, where ATM/ATR phosphorylation targets the Rec114 homolog, which highlights the high degree of conservation of this important regulation.

Globally, the study is well performed, the conclusions are well supported by the experimental data, and both the conservation, but also the specific effect seen during aging, add value to this study and make it of broad interest, in particular to the meiosis and C. elegans communities.

Reviewer #1 (Public Review):

In this study the authors investigate the role of Fam49a and Fam49b in T cell development. The phenotype of the Fam49b KO mice suggests that it may be due to increased negative selection. The experiments are well performed and the results are convincing; however, the study is not extensive and lacks a direct mechanistic element. Additional experiments are recommended to increase the impact of the study.

Reviewer #2 (Public Review):

The Family with sequence similarity 49 member B (Fam49b) protein is a newly discovered negative regulator of TCR signaling that has been shown to suppress Rac-1 activity in vitro in cultured T cell lines. However, the contribution of Fam49b to thymic development of T cells is unknown. Since TCR signaling strength controls thymocyte development, Park et al. hypothesized that Fam49b could be critical for thymocyte development in vivo and investigated this by generating a novel knockout mouse line. They show that Fam49b is dispensable for positive selection but is required to prevent overly robust elimination of thymocytes at the negative selection stage, thereby identifying Fam49b as a critical regulator of negative selection. The conclusions of this paper are mostly well supported by data, but some aspects of TCR signaling and data analysis need to be clarified and extended. Besides, this study reveals a novel function of Fam49b as being required for optimal thymic development of lymphocytes and this could be relevant for T cell central tolerance and susceptibility to autoimmunity.

Reviewer #3 (Public Review):

Using Crisp/Cas9 approaches, the authors generated Fam49a and Fam49b knockout mice. An overall analysis of T cells in Fam49a KO mice indicated no major alterations of thymic and peripheral T cell subsets and this strain was not further analyzed. In contrast, Fam49b knockout mice showed reduced numbers of SP thymocytes and peripheral T cells. A further analysis of Fam49b KO thymocytes subsets was performed and the authors analyzed the expression of markers such as TCRb, CD69, CD5, CCR7 as well as cleaved Caspase 3 expression. In addition, alterations in intestinal T cell subsets were observed. Based on the analysis, the authors conclude that Fam49b KO thymocytes undergo excessive negative selection and that Fam49b dampens TCR signal strength. The manuscript is well written and the data are clearly presented. However, the conclusion that there is excessive negative selection should be strengthened by the inclusion of additional experiments.

Reviewer 1 (Public Review):

In this work Allen and colleagues set out to dissect the mechanisms of protein transport through the Sec machinery. The authors implement the NanoLuc assays, MD simulations, and bioinformatics to sketch out a conceptually interesting hypothesis on protein secretion. The findings suggest that pre-proteins move through the Sec machinery by diffusion and that transport of arginine residues accounts for most of the transport time. The conclusions are supported by the data. Overall, this is an interesting paper on a fundamental topic that will be of interest to a wide audience.

Reviewer 2 (Public Review):

In this study, authors tested the effects of various sequence features of a client preprotein (e.g., the charges, hydrophobicity, helix propensity, and residue sizes) as well as of the proton-motive force on the SecA/SecYEG-mediated protein transport. The experimental designs are elegant, and the results are interesting. Especially, the role of Arg residues in decelerating protein transport is notable. Authors corroborate on explaining the negligible impact of Lys on the rate. The explanation based on MD simulation is also reasonable. Using the sophiscated kinetic model and analysis, authors propose a detailed protein transport mechanism supporting the ratcheting model, which is impressive.

Reviewer 3 (Public Review):

The Sec system transports proteins across bacterial membranes and has relevance to myriad biological processes. A common problem in biochemistry is the lack of a good assay for the particular activity of interest and this affected the study of Sec. However, recently, the Collinson lab developed a high-resolution assay for transport, using a split NanoLuc luciferase, where '11S' (the majority of the NanoLuc enzyme) has a small ß-strand removed ('pep86'). When this pep86 is complemented back, the enzyme is active and a luminescence signal is generated allowing transport to be monitored. Previously, this system was used to perform a high-resolution kinetic analysis, a major first for the field, which set the scene for this work.

Here, the authors have built upon the split NanoLuc system to explore the details of the interaction between the charge of the protein substrate and kinetics of transfer. To do this, they created a model substrate using three tandem pSpy domains, the pep86 signal, and the signal sequence. They then modified the sequence of the central domain to mutate different amino acids to alter bulkiness and charge and observe that of all the mutations, introducing positive arginine residues slows transport the most. Here, the data are convincing, with clear trends being observed. The authors then modify their previously published kinetic modelling to determine step size. They find that for nearly all of the variants, their model predicts about the same number of steps apart from the arginine-containing one, where the model breaks down and does not provide a unique solution. The authors postulate that of the two positive amino acids, arginine presents the most difficulty in transport as its pKa is much too high to be deprotonated before transport, whereas lysine may be as the pKa will be lowered in the hydrophobic interior of the translocon. This idea is both intuitive and supported with careful molecular dynamics simulations. The observation that the immutable positive charge on arginines presents intrinsic difficulties to transport is further supported by a bioinformatics analysis by comparing Tat vs Sec mediated transport and by comparing Bacillus halodurans with Bacillus subtilis; as these organisms grow at different pH values their ΔpH should be radically different, which should affect the ease of transporting arginine residues. Finally, inverted vesicle systems are used to measure the transport of the model substrates while a PMF is established through ATP hydrolysis. Here the data are less clear cut and and it appears that PMF has a major effect on transport depending on pre-protein hydrophobicity (rather than depending on charged residues as one might expect). This is an intriguing finding to end the paper on and I'm sure is something that the authors can build upon in the future.

Overall, the paper is a seminal advance in understanding the details of protein transport through the Sec translocon. The authors are careful to caveat their statements throughout so as to not overstate the strength of their findings and honestly report both clear and 'messy' observations.

Reviewer #1 (Public Review):

Antibiotics of different classes often have the unfortunate property of inhibiting mitochondrial protein synthesis in humans. To overcome these off-target activities, it would be helpful to have high-resolution views of how these antibiotics bind the mitochondrial ribosome to enable future medicinal chemistry efforts. However, to date, no such high-resolution views are available, even for the analogous bacterial system. In this manuscript, Itoh et al. present a cryo-EM reconstruction of the human mitochondrial small ribosomal subunit bound to streptomycin, at a local resolution of ~2.2 Å. With this resolution, the authors are able to define the binding interactions between the compound and the binding pocket in the ribosome. Notably, the antibiotic seems to adopt the gem-diol form in the streptose ring, and also to involve a nearby magnesium ion with one inner-sphere coordination to an rRNA phosphate group. This structure provides an important advance in our understanding of how streptomycin binds the mitochondrial ribosome that could aid future efforts to improve the therapeutic properties of streptomycin derivatives. There are a few issues the authors should address, however, as described below.

First, the authors suggest that adding streptomycin to the human cells prior to ribosome preparation ensures the complex is more physiological. It's not clear that they would have observed a different result had they purified mitochondrial small subunits and subsequently added the antibiotic in vitro. Rather, the more important point the authors could make is that the off-rate of the antibiotic must be rather slow for it to remain bound through the purification steps. (It would help to know just how much washing was done and with what volumes and times.) Is there evidence for slow off-rates, for example from wash-out experiments?

Second, the authors present an "unbound" structure in Figure 1c on the right. Where are the data for this structure? It also needs to be described and deposited.

Third, it is quite puzzling how the central streptose ring could be so flexible, given that the first and third rings are so well constrained by the binding pocket. Could the authors tell us whether it is possible that there is a mixture of aldehyde and gem-diol? It would help to see the aldehyde form modeled and fit to the density. The authors should present correlations for density fit as well.

Reviewer #2 (Public Review):

The overall aim of this work was to understand binding of the aminoglycoside antibiotic streptomycin to the human mitochondrial ribosome. This is an undesirable off-target effect that occurs presumably because the mitochondrial ribosome remains quite similar to the bacterial ribosome from which it evolved. Understanding this binding and comparing it to streptomycin-bound ribosomes from target bacteria may allow for the development of new antibiotic variants without this problematic off-target binding.

To develop this understanding the authors chose to use cryogenic electron microscopy (cryo-EM), the now-standard method for determining the structures of large macromolecular complexes like this. Of particular note, rather than taking the more common route of isolating the ribosomes and then adding the ligand, they instead chose to treat actively-growing human cells with streptomycin prior to isolating the ribosomes. The fact that the streptomycin appears (exceedingly) clearly in the final map despite the fact that none was included in any of the purification steps is very strong evidence that this is a physiologically-relevant result. One obvious (and somewhat sobering) question this raises for me: what are the implications of this for the vast body of in vitro mammalian cell culture research carried out with a cocktail of penicillin and streptomycin added to the media for infection control?

The resolution of the final map (2.23Å) represents a dramatic improvement over the previous-best 2.97Å map of the mitochondrial ribosome small subunit, and to the authors' credit, it is clear that a great deal of effort has gone into high-quality modelling throughout. I am very happy to see that since it remains sadly common for some authors to focus their efforts only on the region of their immediate interest. In this case, the focal site is the bound streptomycin - which is clear and unambiguous, and in my opinion fully supports the authors' conclusions regarding the details of binding, most notable of which is that the aldehyde group found on one of the streptomycin sugar residues becomes hydrated to the geminal diol form (two hydroxyl groups bound to the same carbon) prior to binding the ribosome.

Reviewer #3 (Public Review):

In this manuscript, the authors set out to determine the structural basis of off-target binding to the human mitoribosome of the anti-tuberculosis antibiotic streptomycin using the latest advances in cryo-electron microscopy. The maps they have generated are of very high quality, enabling them to visualize streptomycin bound to the shoulder of the small subunit with the coordinated waters and magnesium molecules at an impressive overall nominal resolution of 2.3 A. The resolution of the maps represents a significant improvement compared to the previously published X-ray structures of streptomycin bound to the bacterial ribosome, allowing the authors to propose an improved model for the binding mode of streptomycin. In particular, the authors find that the density for a previously modeled aldehyde group in the bacterial ribosome-streptomycin complexes is poorly defined. In contrast to previous work, the authors interpret the density as consistent with a hydrated aldehyde, a model that is supported by previously published solution NMR spectroscopic analysis.

As streptomycin is still widely used for anti-TB therapy, this structure and the methodological advances involved in achieving such high-quality maps will be of interest to investigators in the mitochondrial biology field and more widely within the community of researchers interested in structure-guided drug design applied to infectious disease to minimize treatment-related toxicities. The data clearly reveal how streptomycin binds to the mitochondrial small subunit but some aspects of the study require further clarification.

1) The precise source of the streptomycin used in the study is not currently stated in the text. It is not clear if the compound was subjected to detailed mass spectrometry analysis to validate the composition assumed by the authors and to identify any potential chemical heterogeneity.

2) Although the overall density strongly supports the presence of bound streptomycin, the map does not convincingly support the proposed hydrated aldehyde within the streptomycin streptose moiety. While the claim that the poor density may reflect increased mobility of the hydrated aldehyde as the authors suggest is completely reasonable (and is supported by previously published NMR data), this would need to be independently validated given the ambiguity in the map. A structure of streptomycin added in vitro to the mitochondrial small subunit would resolve whether the added molecule is indeed modified by the cellular milieu. While the authors suggest that adding the streptomycin to cultured cells rather than in vitro is an advantage of the work, the disadvantage is that it generates uncertainty that has not been resolved by the cryo-EM map alone about the chemical nature of the bound compound. The density might also be consistent with a potassium ion, for example.

3) Given the high quality of the map, it is surprising that there is no discussion about the remainder of the mitochondrial small subunit, including ligands/spermine listed in Table 1. What is the rationale for GMPPNP and ATP, which are listed as ligands in Table 1?

Reviewer #1 (Public Review):

The authors sought to achieve an understanding of the cell types present in the ovaries of a young mature zebrafish adult, define the set of genes each cell type expresses, to identify cell types in histological sections, and to identify function for two genes by induced mutations.

The work has several major strengths. The methodologies were appropriate for the questions posed. The results were analyzed in depth, looking at the expression patterns in the scRNA-seq data for a substantial number of genes with a focus on those genes' likely functions. The single molecule RNA in situ hybridization experiments tied the dots-on-a-sheet expressions to cells in actual animals. The identification of subpopulations of cell types is useful to understand development moving forward. The estrogen story is nice as is straightening out the foxl2 paralog naming system. The work has few detracting weaknesses.

The authors' datasets and analysis achieved their stated aims and results support conclusions.

The work will have a positive impact on the field, having identified many gene sets to be explored by the community by mutation and other types of experiments.

Reviewer #2 (Public Review):

The goal of this study was to describe the cell types and hierarchies present in the juvenile zebrafish ovary using single-cell RNA sequencing. The authors were successful in describing unexpected complexities in the various cell types of the ovaries including follicle cells, theca cells, and interstitial stromal cells. For germ cells, they were able to provide molecular profiles of pre-meiotic, meiotic and early-stage oocytes. For the ovarian mesenchyme, they describe follicle cells and pre-follicle cell subtypes, theca cell subpopulations, and other stromal cell subpopulations, including a putative stromal progenitor. Many of these cell subtypes were validated by their expression of molecular markers and their histologies. The authors also produced reverse genetic evidence of the role of newly identified genes in the regulation of progenitor functions in germ cells stem cells and pre-follicle cells. Finally, they used steroid biosynthesis as a test case for the utility of this cell atlas in elucidating cellular and molecular functions.

Major strengths:

The hypothesis that Foxl2l is a marker of oocyte progenitors was well supported by histological evidence and the localization of nanos2 and rec8a. Furthermore, the sex ratio skewing observed in the homozygous knock-in allele recapitulates its role in medaka as a necessary transcription factor to commit to female germ cell differentiation.

Similarly, disruption of the wnt9b gene perturbs sex determination, and partially phenocopies previous findings of wnt4a, suggesting this pathway is important in early follicle cell specification.

Another important finding is the source of E2 biosynthesis in the zebrafish which appears to favor Hsd17b1 to catalyze the E2 synthesis, which is expressed primarily in follicle cells. This mirrors the 2 cell hypothesis of E2 production in mammalian follicles but does not rely on cyp19a1. This is supported both by molecular evidence and careful analyses of previously uncharacterized follicular structures.

Weaknesses:

The identification of a putative stromal progenitor is one of the more exciting findings of the study. Yet their molecular signature and spatial position within the ovary were not validated.

Likewise, the presence of distinct subpopulations of theca is not fully explored. It is unclear whether the subsets are associated with different maturation stages of the follicle, and thus might represent immature versus differentiated theca.

Likely impact:

This dataset identified novel cell types and molecular pathways which will likely be the subject of fertile investigation. Furthermore, many of the genes identified herein are likely to be developmentally relevant and should be the subject of further gene knockout analysis.

The data provides strong evidence of conservation of function of key gonadal sex-determining genes, including those of the Wnt and Foxl2 family, which have orthologs in mammals and fish. Therefore, this new understanding of ovarian cells provides a framework for using zebrafish as a model for ovarian development and sexual differentiation in humans, and perhaps for the study of developmental disorders and diseases.

Reviewer #3 (Public Review):

This scRNAseq study provides definitive markers for various ovarian cell types in the zebrafish and new cell sub-types, as well as specific ovarian germ cell stages and early oocyte developmental stages, which will be valuable for future functional studies. Importantly, this study makes clear the strongly conserved nature of oogenesis between zebrafish and mammals, demonstrating conserved expression signatures for various cell types. Beautiful FISH analysis of definitive marker genes of particular subpopulations was used to determine where the subpopulation resides in the ovary. The gene motif enrichment analysis to identify putative transcriptional regulators of gene modules is quite compelling and valuable in providing testable frameworks of regulatory pathways. A further strength is the functional analysis of two genes that were mutated to test for predicted functions.

Reviewer #1 (Public Review):

Since Kimura, it is has been clear that most of the DNA sequence polymorphism in the population must, in general, be selectively neutral (or effectively so). At the same time, it has been clear that selection will affect the pattern of polymorphism through linkage. It will leave footprints. In particular, levels of polymorphism can be reduced by the constant elimination of deleterious mutations by purifying selection - and also by repeated adaptive substitutions ("selective sweeps"). Both could explain the observed pattern of less polymorphism in gene-dense regions. But because they are very different mechanisms, it is of interest to figure out which one is most important. Flipping the argument around could help us understand how prevalent and strong each is.

This paper is a rigorous, state-of-the-art attempt to do this in the human genome. Using the latest data and computational methods, the paper convincingly argues that purifying selection must be the dominant force - indeed, it provides a surprisingly good predictive, mechanistic model, which is rare in population genetics. New insights include the major contribution of purifying selection on non-coding regions, which is evident from sequence conservation. The paper, which is exceptionally well written, should be of interest to anyone interested in molecular population genetics.

Reviewer #2 (Public Review):

Murphy et al. extend previous models of linked selection to evaluate diversity in the human genome. They find that patterns of diversity along the genome can be largely explained by models of background selection without the need to include selective sweeps or even much functional information beyond the presence of evolutionary constraints.

This paper is excellent. One of the papers I have most enjoyed reading in some time. Many of the concerns or questions I had while reading were immediately answered by the authors, who have addressed almost every concern I can think of as well as many many more. In multiple places, while reading the supplement I found myself thinking "wow that's a good point I never thought of that but I'm glad they did". The figures are clear and well explained. The supplement is a veritable trove of thoughtful background, careful consideration of caveats and concerns, and well-reasoned arguments for the choices the authors made.

Given the above, it may not be surprising that my concerns with the paper are relatively few:

The authors address polygenic adaptation, but I wouldn't mind additional discussion of how polygenic adaptation to a rapidly fluctuating optimum might or might not be captured by the background selection model.

I would like to see a bit better introduction for readers not already steeped in linked selection. Inclusion of the basic equations from Hudson and Kaplan or Nordborg might go a long way to helping an average reader (who may not have the stamina for 80 pages of supplement) to nonetheless understand the basic parameters and model that the authors build off of.

Similarly, the last bit before the conclusions could be expanded a bit. How do the authors propose empiricists could use their results most effectively for e.g. demographic estimation? What are other areas/uses/implications of the results for other evolutionary genetic work?

Reviewer #3 (Public Review):

Murphy et al. further develop the linked selection model of Elyashiv et al. (2016) and apply it to human genetic variation data. This model is itself an extension of the McVicker et al. (2009) paper, which developed a statistical inference method around classic background selection (BGS) theory (Hudson and Kaplan, 1995, Nordborg et al., 1996). These methods fit a composite likelihood model to diversity data along the chromosome, where the level of diversity is reduced by a local factor from some initial "neutral" level π0 down to observed levels. The level of reduction is determined by a combination of both BGS and the expected reduction around substitutions due to a sweep (though the authors state that these models are robust to partial and soft sweeps). The expected reduction factor is a function of local recombination rates and genomic annotation (such as exonic and phylogenetically conserved sequences), as well as the selection parameters (i.e. mutation rates and selection coefficients for different annotation classes).

Overall, this work is a nice addition to an important line of work using models of linked selection to differentiate selection processes. The authors find that positive selection around substitutions explains little of the variation in diversity levels across the genome, whereas a background selection model can explain up to 80% of the variance in diversity. Additionally, their model seems to have solved a mystery of the McVicker et al. (2009) paper: why the estimated deleterious mutation rate was unreasonably high. Throughout the paper, the authors are careful not only in their methodology but also in their interpretation of the results. For example, when interpreting the good fit of the BGS model, the authors correctly point out that stabilizing selection on a polygenic trait can also lead to BGS-like reductions.

Furthermore, the authors have carefully chosen their model's exogenous parameters to avoid circularity. The concern here is that if the input data into the model - in particular the recombination maps and segments liked to be conserved - are estimated or identified using signals in genetic variation, the model's good fit to diversity may be spurious. For example, often recombination maps are estimated from linkage disequilibrium (LD) data which is itself obtained from variation along the chromosome. Murphy et al. use a recombination map based on ancestry switches in African Americans which should prevent "information leakage" between the recombination map and the BGS model from leading to spuriously good fits. Likewise, the authors use phylogenetic conservation maps rather than those estimated from diversity reductions (such as McVicker et al.'s B maps) to avoid circularity between the conserved annotation track and diversity levels being modeled. Additionally, the authors have carefully assessed and modified the original McVicker et al. algorithm, reducing relative error (Figure A2).

One could raise the concern that non-equilibrium demography confounds their results, but the authors have a very nice analysis in Section 7 of the supplementary material showing that their estimates are remarkably stable when the model is fit separately in different human populations (Figure A35). Supporting previous work that emphasizes the dependence between BGS and demography, the authors find evidence of such an interaction with a clever decomposition of variance approach (Figure A37). The consistency of BGS estimates across populations (e.g. Figures A35 and A36) is an additional strong bit of evidence that BGS is indeed shaping patterns of diversity; readers would benefit if some of these results were discussed in the main text.

I have three major concerns about this work. First, it's unclear how accurate the selection coefficient estimates are given the non-equilibrium demography of humans (pre-Out of Africa split, and thus not addressed by the separate population analyses). The authors do not make a big point about the selection coefficient estimates in the main section of the paper, so I don't find this to be a big problem. Still, some mention of this issue might be helpful to readers trying to interpret the results presented in the supplementary text.

Second, I'm curious whether the composite likelihood BGS model could overfit any variance along the chromosome - even neutral variance. At some level, the composite likelihood approach may behave like a sort of smoothing algorithm, albeit with a functional form and parameters of a BGS model. The fact that there is information sharing across different regions with the same annotation class should in principle prevent overfitting to local noise. Still, there are two ways I think to address this overfitting concern. First, a negative neutral control could help - how much variation in diversity along the chromosome can this model explain in a purely neutral simulation? I imagine very little, likely less than 5%, but I think this paper would be much stronger with the addition of a negative control like this. Second, I think the main text should include the R2 values from out-sample predictions, rather than just the R2 estimates from the model fit on the entire data. For example, one could fit the model on 20 chromosomes, use the estimated θΒ parameters to predict variation on the remaining two. The authors do a sort of leave-one-out validation at the window level (Figure A31); however, this may not be robust to linkage disequilibrium between adjacent windows in the way leaving out an entire chromosome would be.

Finally, I feel like this paper would be stronger with realistic forward simulations. The deterministic simulations described in the supplementary materials show the implementation of the model is correct, but it's an exact simulation under the model - and thus not testing the accuracy of the model itself against realistic forward simulations. However, this is a sizable task and efforts to add selection to projects like Standard PopSim are ongoing.

Reviewer #1 (Public Review):

Protein targeting of precursor proteins to the mitochondrial surface is poorly understood. Drwesh and colleagues characterized cytosolic factors that target mitochondrial signal-anchored proteins. Using yeast translational extracts, they identified Hsp70 and Hsp90 chaperones and their co-chaperones such as Ydj1 and Sis1 as important binding partners. The transmembrane domain of the signal anchored protein is critical for the binding of the chaperones. The authors used a combination of in vitro binding assays and import experiments into mitochondria to characterize the interaction of the chaperones to the precursor proteins and their subsequent transfer to the TOM receptors. Overall, the reported findings are highly interesting and an important contribution to the field. The study provides important insights into the cytosolic network that controls protein targeting to mitochondria. The presented data are of high quality and convincing.

Reviewer #2 (Public Review):

Drwesh and colleagues used an array of in vitro, in organello and in vivo approaches to study the early steps of mitochondrial signal-anchored (SA) membrane protein biogenesis after translation and before membrane insertion. The authors chose four distinct yeast mitochondrial outer membrane SA proteins (Tom20, Tom70, Msp1 and the outer membrane form of Mcr1, Mcr1mom) and analyzed their recognition by cytosolic chaperones and cochaperones from the Hsp70 and Hsp40 families that bind translated precursors, preventing their aggregation and keeping precursors in an import-competent state. Furthermore, they analyzed precursor protein interactions with outer mitochondrial membrane receptors, therefore covering the two major cytosolic steps required for the biogenesis of mitochondrial SA proteins.

The authors could clearly show that SA protein recognition via co-/chaperones occurs via interactions with their transmembrane segments. Using co-immunoprecipitation of newly-translated HA-tagged precursors in yeast extract the authors found all analyzed SA proteins to interact with chaperones from the Hsp70 (Ssa1/2) and co-chaperones from the Hsp40 (Ydj1, Sis1, Djp1) families. Further mass spectrometry analysis of the Mcr1 and Msp1 eluates showed their interaction with a broader array of chaperones from the Hsp70 family.

Downregulation of Ydj1 and Sis1 co-chaperones from the Hsp40 family showed an effect on steady-state levels and import defects for SA proteins. Interestingly, depletion of both co-chaperones showed the increase in SA precursor binding to Hsp104 and Hsp26 chaperones that prevent protein aggregation, suggesting that co-chaperones from the Hsp40 family, even if they do not directly impact steady-state levels or import of all tested SA candidates, they do prevent their aggregation.<br /> In vitro fluorescence anisotropy experiments demonstrated the interaction of the transmembrane segment of some SA candidates for both Ssa1 and Sis1, with a higher affinity towards the chaperone as for the co-chaperone. Interestingly, using the same approach and sequential addition of co-/chaperones and membrane receptor domains, they showed that precursors can be transferred in vitro from the co-chaperone, to the chaperone and to the receptors, therefore reconstituting the required steps in early biogenesis of SA proteins. Affinity measurements go in accordance with this sequential precursor transfer.

Finally, the authors show that even if the studied membrane receptors are cleaved or blocked, SA proteins can still be imported into mitochondria in organello to different extents, suggesting that the cytosolic chaperone system is even more important SA anchored protein biogenesis compared to the import receptors.

Overall, the manuscript is the first in-depth characterization of the complex transfer of mitochondrial signal-anchored proteins by cytosolic co-/chaperones to the mitochondrial membrane receptors, suggesting that cells possess redundant chaperone and receptor machineries to fulfill this task. This set of distinct approaches characterizes the cytosolic interactions required to efficiently transfer signal-anchored membrane proteins from co-/chaperones to the membrane receptors.

Reviewer #3 (Public Review):

The authors aimed to investigate the network of interactions in the cytosol between a number of chaperones and SA targeted proteins in the mitochondrial OM. They also wanted to clarify the mechanism of insertion at the OM and the role of the main receptors in this process. They have managed to provide a wealth of interesting data on a mechanism that is very elaborate as it involves a lot of redundancy between the different chaperones. One strong point of their work is that they go beyond the typical analysis of interactions by biochemical approaches and provide some quantitative data that support a model for a cascade of interactions. This was missing in previous studies and is a notion that will likely be taken up in future studies as well. The impact of the work is substantial and represents an important step forward in our understanding of the mitochondrial protein targeting processes.

Reviewer #1 (Public Review):

This paper describes an analysis of the role of K63 ubiquitylation of ARTs proteins by Rsp5. ARTs are adaptors for Rsp5 that link it to recruitment to membranes including the plasma membrane where the ubiquitylation activity of Rsp5 promotes cargo endocytosis. Rsp5 di-ubiquitylates several ARTs and this enhances Rsp5 recruitment to membranes and is required for cargo ubiquitylation. Interestingly, the HECT domain exosite blocks the di-K63 ubiquitin from removal by DUBs. When the exosite interaction is not present, di-K63 is removed and ARTs become substrates for K48 ubiquitylation and proteasomal degradation. These studies reveal a detailed mechanism of ARTs modification that couples Rsp5 activity to activation of ARTs for cargo import. The experiments are generally of very high quality and robustly demonstrated.

Reviewer #2 (Public Review):

In their manuscript Zhu, et. al. describe an intriguing mechanism by which the activity Arrestin-Related Trafficking adaptors (ARTs) is modulated by cycles of Rsp5-dependent ubiquitination and Ubp2-dependent de-ubiquitination. Rsp5 is a NEDD4 family E3 ligase, which binds to target proteins via PY motifs. In some cases, PY-containing adaptor proteins are required to bridge the interaction between Rsp5 and its targets. The ART proteins in yeast are a family of such adaptor proteins, and regulation of their activity is crucial for maintaining cellular homeostasis, particularly in response to environmental changes. The authors demonstrate that Art1, Art4, and Art5 are di-ubiquitinated by a K63 linkage in vivo and that this modification does not alter the ability of the ARTs to recognize their target receptor, but rather enhances the ART-Rsp5 interaction, via ubiquitin binding to Rsp5's exosite. Furthermore, they show that loss of the K63 ubiquitination of the arrestin proteins, facilitated by Ubp2, allows for K48 mediated ubiquitination and recycling of the arrestin by proteasomal degradation.

Reviewer #3 (Public Review):

Prior studies including from Emr's lab showed that ubiquitination of adaptor proteins is required for the recruitment of Rsp5/Nedd4 HECT ubiquitin ligase to regulate endocytosis and degradation of cell surface cargo proteins. In this manuscript, Zhu et al. investigated the mechanism by which the ubiquitinated adaptor proteins, Art1, Art4 and Art5, could facilitate the ubiquitin ligase function of Rsp5. They found that these adaptor proteins are predominantly modified with di-ubiquitin on a lysine site in cells. K63R-ubiquitin analysis indicated that the di-ubiquitin is K63-linked and this di-ubiquitin modification is dependent on the presence of PY motif, important for interaction with Rsp5. This is consistent with prior studies showing that Rsp5 functions to ubiquitinate Art adaptor protein. They demonstrated that Art1 and Art5 interaction with their respective cargo proteins, Mup1 and Itr1, is not affected by di-ubiquitin modification, but ubiquitination deficient mutants of Art1 and Art5 had defects in sorting and the ubiquitination of cargo protein after nutrient stimulation. They demonstrated that recruitment of Rsp5 to plasma membrane requires the ubiquitinated Art adaptor proteins and the cargo substrates. To investigate the relevance of K63-linked di-ubiquitin modification on Art protein, they investigated Rsp5 exosite, which is known to bind mono-ubiquitin. They showed that Rsp5 exosite has a higher binding affinity for K63-di-ubiquitin and linear-di-ubiquitin than mono-ubiquitin and K48-di-ubiquitin. Moreover, I44A-substitution in the distal K63-di-ubiquitin reduces the binding affinity suggesting that both ubiquitin moieties contribute to Rsp5 binding. By generating Any1 (adaptor protein from S. pombe) modified with K63-di-ubiquitin, they showed that di-ubiquitin enhances the binding affinity with Pub1 (S. pombe Rsp5) by ~8-fold. These data allow the authors to propose a model where di-ubiquitin modification of adaptor protein enhances the binding affinity for Rsp5 leading to the recruitment of Rsp5 to cargo proteins at the plasma membrane to facilitate Rsp5-mediated cargo ubiquitination and sorting. Lastly, the authors showed that ubp2 could reverse K63 ubiquitination of adaptor protein to allow K48-ubiquitination of adaptor protein.

Overall the manuscript is clear. The cell-based data showed the importance of Art adaptor proteins ubiquitination in Rsp5-mediated cargo ubiquitination and sorting. The Art adaptor proteins predominantly exist in the di-ubiquitinated form and K63R-ubiquitin mutant suggested that the di-ubiquitin is linked via K63. These data point toward a role of K63-di-ubiquitin modified Art in the regulation of Rsp5. Addressing K63-di-ubiquitinated Art in the regulation of Rsp5 in cells might be challenging, but the authors demonstrated that K63-di-ubiquitin modified Any1 exhibited a higher binding affinity for Pub1 and this is dependent on Rsp5 exosite. Together these data support the proposed model. The proposed mechanism of ubp2 in reversing K63-ubiquitination in the adaptor protein to enable K48-ubiquitination of Art protein is not well developed here. Much of this hypothesis was derived using a Rsp5 mutant that is defective in exosite ubiquitin binding.

Reviewer #1 (Public Review):

The paper is very well written, the question is interesting, and the analyses are innovative. However, I do have concerns about the overall approach. My main concern is about looking at asymmetries in the low dimensional representation of connectivity. A secondary concern has to do with looking at the parcellated connectome. I explain these concerns in succession below.

The first concern is to me quite a fundamental issue: looking at connectivity in a low dimensional space, that of the laplacian eigenvectors. There are two issues with this. The first one, which is less important than the second, is that the authors have a reference embedding to which they align other embeddings using a procrustes method with no scaling. While the 3D embedding is still optimally representing the connectivity (because distances don't change under rotations), we can no longer look at one axis at a time, which is what the authors do when they look at G1. In this case, G1 is representative of the connectivity of the reference matrix (LL), but not the others.

But even if the authors only projected their matrices onto a single G1 dimension with no procrustes (and only sign flipping if necessary), there is still a major issue. One implicit assumption of this whole approach is that if there is a change in connectivity somewhere in the original matrix, the same "nodes" of the matrix will change in the embedding. This is not the case. Any change in the original matrix, even if it is a single edge, will affect the positions of all the nodes in the embedding. That is because the embedding optimises a global loss function, not a local one.

To make this point clear, consider the following toy example. Say we have 4 brain regions A,B,C,D. Let us say that we have the following connectivity:

In the Left Hemisphere: A-B-C-D<br /> In the Right Hemisphere: A-B=C-D

So the connection between B and C is twice as strong in the right hemi, and everything else remains the same.

The low dimensional embedding of both will look like this:

Left:<br /> ... A ... B ....... C ... D ...<br /> Right<br /> A... ... ... B ... C ... ... ... D

Note how B,C are closer to each other in the RIGHT, but also that A,D have moved away from each other because the eigenvector has to have norm 1.

So if we were to calculate an asymmetry index, we would say that:

A is higher on the LEFT<br /> B is higher on the RIGHT<br /> C is higher on the LEFT<br /> D is higher on the RIGHT

So we have found asymmetry in all of our regions. But in fact the only thing that has changed is the connection between B and C.

This illustrates the danger of using a global optimisation procedure (like low-dim embedding) to analyse and interpret local changes. One has to be very careful.

My second concern is about interpreting the brain asymmetry as differences in connectivity, as opposed to differences in other things like regional size. The authors use a parcellated approach, where presumably the parcels are left-right symmetric. If one area is actually larger in one hemisphere than in the other, the will manifest itself in the connectivity values. To mitigate this, it may be necessary to align the two hemispheres to each other (maybe using spherical registration) using connectivity prior to applying the parcellation.

Reviewer #2 (Public Review):

Using recently-developed functional gradient techniques, this study explored human brain hemispheric asymmetry. The functional gradient is a hot technique in recent years and has been applied to study brain asymmetries in two papers of 2021. Compared to previous studies, the current study further evaluated the degree of genetic control (heritability) and evolutionary conservation for such gradient asymmetries by using human twin data and monkey's fMRI data. These investigations are of value and do provide interesting data. However, it suffers from a lack of specific hypotheses/questions/motivations underlying all kinds of analyses, and the rich observational or correlational results seem not to offer significant improvement of theoretical understanding about brain asymmetries or functional gradient. In addition, given the limited number of twins in HCP project (for a heritability estimation), the limited number of monkeys (20 monkeys), and the relatively poor quality of monkeys' resting functional MRI data, the results and conclusion should be taken cautiously. Below are major concerns and suggestions.

The gradient from resting-state functional connectome has been frequently used but mainly at the group level. The current study essentially applied the gradient comparison (i.e., gradient score) at the individual level. Biological interpretation for individual gradient score at the parcel level as well as its comparability between individuals and between hemispheres should be resolved. This is the fundamental rationale underlying the whole analyses.

Only the first three gradients are used but why? What about the fourth gradient? Specific theoretical interpretation is needed. At the individual level, is it ensured that the first gradients of all individuals correspond to each other? In this study, it is unclear whether we should or should not care about the G2 and G3. The results of G2 and G3 showed up randomly to some degree.

The intra-hemispheric gradient is institutive. However, it is hard to understand what the inter-hemispheric gradient means. From the data perspective, yes you can do such gradient comparison between the LR and RL connectome but what does this mean? Why should we care about such asymmetry? From the introduction to the discussion, the authors simply showed the data of inter-hemispheric gradients without useful explanation. This issue should be solved.

When aligning intra-hemispheric gradient, choosing averaged LL mode as the reference may introduce systematic bias towards left hemisphere. Such an issue also applies to LR-RL gradient alignment as well as cross-species gradient alignment. This methodological issue should be solved.

The sample size of monkey (i.e., 20) is far less than human subjects (> 1000). Such limitation raises severe concern on the validity of the currently observed gradient asymmetry pattern in the monkey group, as well as the similarity results with human gradient asymmetry pattern. Despite the marginal significance of G1 inter-hemisphere gradient between humans and monkeys, I feel overall there is no convincingly meaningful similarity between these two species. However, the authors' discussion and conclusion are largely based on strong inter-species similarity in such asymmetry. The conclusion of evolutionary conservation for gradient asymmetry, therefore, is not well supported by the results.

Reviewer #1 (Public Review):

I previously highlighted the need for a physiologically relevant cell type, the issue of showing that OCT2 is not only sufficient but also required for activation of Golgi-localized receptor, and a concern about how cytoplasmic dopamine gains access to the Golgi lumen. While the latter concern somewhat remains, this version satisfactorily addresses these issues and I believe the study will be of interest to a large audience.

Reviewer #2 (Public Review):

Puri, Romano, et al. investigated the possibility that, as they and others have demonstrated for adrenergic receptors, functional dopamine D1 receptors can be localized to and activated at Golgi membranes. They use an impressive array of genetically encoded biosensors to visualize the cellular localization of dopamine receptors, their activation, and downstream signaling events. They find that bath application of dopamine, a membrane impermeant ligand for D1 receptors, causes activation of Golgi-localized D1 receptors in HeLa cells and primary cultured striatal neurons, but not in HEK293 cells. They find that dopamine activation of Golgi-localized D1 receptors requires expression of OCT2, a transmembrane catecholamine transporter. In HEK cells, which they show express low levels of OCT2, Golgi-localized D1 receptors can be activated by a membrane-permeant D1 receptor agonist, leading to apparent activation of PKA and contributing to D1-induced increases in cAMP concentration. These findings add to the growing body of evidence that G-protein-coupled receptors for monoamines can be localized to intracellular membranes and activated by their endogenous ligands in a transporter-gated fashion. Intracellular D1 receptors represent an intriguing novel mechanism by which dopamine may influence neuronal activity, and which may contribute to the physiological and behavioral actions of this catecholamine transmitter.

The conclusions of this manuscript are generally supported by the data. The authors have demonstrated that D1 receptors can be localized to, and activated at, the Golgi apparatus, and that activation of these intracellular receptors requires transmembrane catecholamine transporters. Overall, the work contributes to the growing body of evidence that G-protein-coupled receptors can be activated not only at the plasma membrane, but also from intracellular membranes. This has been demonstrated repeatedly for beta-adrenergic receptors but has never been demonstrated for dopamine receptors. While the authors do not describe physiological consequences of Golgi D1 activation in striatal neurons, the demonstration that they occur and are activated in striatal neurons has profound implications for our understanding of the mechanisms by which dopamine and drugs of abuse influence striatal function.

Strengths:

The use of multiple genetically-encoded fluorescent sensors has allowed the authors to build a strong base of evidence that dopamine D1 receptors can be localized to the Golgi and that they can be shifted to their active conformation by D1 ligands. They also provide evidence that activation of Golgi-localized receptors contributes to D1-induced increases in cAMP and leads to activation of PKA. Thus, Golgi D1 receptors appear to be fully functional Gs-coupled receptors.

The authors make very effective use of the differences in Golgi D1 activation between HeLa and HEK cells to illustrate the potential role of OCT2 in dopamine-induced activation of intracellular receptors. They make effective use of OCT inhibitors and genetic overexpression of OCT2 to demonstrate that dopamine activation of Golgi D1 receptors requires the expression of this catecholamine transporter. The further use of the membrane permeant D1 agonist both confirms the identity of the receptors and demonstrates the necessity of transporter expression and function for dopamine-induced activation.

The authors have provided evidence that Golgi-localized D1 receptors also occur in cultured mouse striatal neurons, which suggests that this signaling system may contribute to the powerful actions of dopamine in these neurons. This would represent a highly novel potential mechanism underlying dopamine-induced regulation of neuronal physiology and behavior. The authors demonstrate, using immunofluorescence that endogenous D1 receptors are localized to Golgi in cultured neurons. It is unclear why, in their subsequent imaging experiments, they used exogenously expressed SNAP-GFP tagged D1 receptors.

Weaknesses:<br /> Most of the data are from cell lines exogenously expressing D1 receptors. The Golgi localization and OCT2-dependent activation are demonstrated in cultured murine striatal neurons, but downstream signaling or physiological responses are only described in HeLa and/or HEK cells.

The selectivity of the transport inhibitors is overstated. Corticosterone is described as an OCT3 inhibitor when it also inhibits OCT2. Imipramine is described as an OCT2 inhibitor when it also inhibits OCT1, OCT3, and the plasma membrane monoamine transporter (PMAT). Given that only OCT2 expression is quantified in any of the cells under study, a clear description of the relative potencies of these inhibitors at the other transporters is necessary to justify the definitive conclusions the authors make about the exclusive role of OCT2. The authors cite work demonstrating OCT localization to intracellular membranes, including nuclear and Golgi membranes. This work focused exclusively on OCT3. This must be clearly stated.

There are instances in which the conclusions made by the authors are not fully justified by the data. The authors state that OCT2 expression is "negligible" in hippocampal tissue, but there is a clear OCT2-immunoreactive band in the western blot. They state that HEK293T cells "do not express OCT2", but there is a clear OCT2-immunoreactive band in their western blot. Also regarding the western blot data: The authors describe primary murine striatal MSNs where, "OCT2 is expressed at high levels." The data they refer to describe OCT2 expression in bulk striatal tissue which, while it does include MSNs, also includes other neuronal types, glial cells, and vascular tissue.

There was no specific measurement of OCT2 expression specifically in MSNs, so the statement overstates the findings of the western blot.

Reviewer #3 (Public Review):

This study repurposes Nb6B9, a nanobody originally used to stabilize adrenergic receptors in an active conformation, to detect active D1 Dopamine Receptors at different regions of the cell. Dopamine is known to be transported by the OCT2 transporter. The authors observe that HeLa cells and cultured primary neurons which have endogenous OCT2, recruit Nb6B9 to exogenously expressed D1DR in the Golgi, but HEK293 cells which do not have endogenous OCT2 do not. HeLa cells also recruit other biosensors that read out activated Gs protein and PKA. Expressing OCT2 in HEK293 cells causes recruitment of Nb6B9 in response to dopamine.

GPCR signaling from Golgi is an emerging phenomenon that is still not well understood. Recent studies have reported Golgi activation of adrenergic receptors by norepinephrine and opioid receptors by synthetic opioids. The main advance of this study is that it adds dopamine receptors to this list. As with these previous papers, this study takes advantage of biosensors that are still relatively new in the field.

The main weakness is the physiological relevance of the observation. As it stands, how much dopamine gets into the Golgi and whether it activates endogenous D1DR are not clear.

The authors use 10µM dopamine in some assays and 10nM in others, which complicates the interpretation. Gründemann 1998, referenced by the authors, have provided rates of transport of dopamine by OCT2, which the authors could use to estimate how much dopamine will get into the Golgi over time. The authors can match a dose response of dopamine to these estimates. This is also important as Nb6B9 recruitment of HEK293 cells seem to increase over ~1000 sec, which is comparable to Nb80 recruitment to B1AR by norepinephrine (~20 min).

Another concern is the specificity of some of the reagents used. For example, a high dose of imipramine is used to block OCT2. Imipramine acts on many other targets, including monoamine uptake and D2 dopamine receptor. Genetically depleting OCT2 in neurons, or at least in HeLa cells, is critical to show that OCT2 is required for Golgi activation.

Repurposing Nb6B9 to detect D1DR is clever. But it also raises concerns about the specificity of the Nb6B9. Does it bind other catecholamine receptors that are in neurons, which could be cross-activated by dopamine? Further, Nb6B9 was originally designed to stabilize an active form of the receptor. The effects could be a due to Nb6B9 expression stabilizing active D1DR.

The miniGs and Nb37 experiments in Supplemental Figures 3 and 4 are also important in this regard, but they are not convincing. Nb37 and miniGs shows much weaker recruitment, which suggests that Nb6B9 might be changing receptor sensitivity. A dose response will help here. Also, it is critical to know how recruitment of all these sensors compare to positive control (e.g., B1AR with NE) and negative control (e.g., opioid receptors) for this experiment and for Fig 1.

The localization of the endogenous D1DR in the Golgi of striatal neurons and the activation by DA is a critical part of the paper. However it is missing controls. GPCR antibodies are notoriously bad, and this commercial D1DR antibody is recommended only for immunoblotting. The authors need to confirm that this antibody is specific.

Overall, this is an interesting paper that presents a new observation that D1DR could be activated on the Golgi in neurons. The paper will be significantly strengthened by improving the physiological relevance and being rigorous in providing controls.

Reviewer #1 (Public Review):

This is an interesting study demonstrating the application of deep learning to model microbiome dynamics of the human gut community, improving on existing approaches (for example regarding scalability). Furthermore, the model is able to better predict microbe-microbe and microbe-metabolite interactions as compared to classical approaches like ODEs or regression. The authors show that their LSTM-based model is able to successfully predict the abundance not only at the final time step but also at intermediate time steps. In general, the authors did a good job in demonstrating the strengths of their proposed approach. The major findings were carefully interpreted and challenged through multiple tests (explainability and sensitivity analysis). As the microbiome is not my primary area of expertise, I cannot comment on the validity of the biological interpretations.

The methods section (machine learning part) is rather short and in my opinion does not provide sufficient details. Since generating deep learning-based models can be rather challenging, it would be valuable to explain how the model was obtained and how the parameter tuning was done. Furthermore, the choice of the LIME and CAM as explanation methods seems arbitrary. It is unclear why these methods are preferable to other methods.

Reviewer #2 (Public Review):

In this study, Baranwal et al demonstrate the use of a long-short memory neural network (LSTM) to predict temporal dynamics of microbial communities and metabolic functions. The key points of the study include:

1. The LSTM model outperforms the standard general Lotka Volterra model in predicting both experimental data and simulated data (from another gLV model).

2. Once trained, the LSTM model allows accelerated prediction of microbial community dynamics.

3. The predictions from LSTM can enable the generation of biological insights by proper analysis.

Overall, this is a very strong paper that represents an important contribution to the field of predicting microbiome dynamics and function using ML. In terms of methodology, I appreciate how the team integrates quantitative measurements, dynamical modeling, and machine learning.

Reviewer #3 (Public Review):

The authors ultimately wish to construct microbiomes with desired functions. To that end they have combined an LSTM model and FF neural network for microbiome and metabolite abundance data that can predict both microbial dynamics and their functional capacity (metabolic potential) over time. Their model is compared to a gLV composite model. Model performance is compared on synthetic data and real data. Sensitivity analysis was performed on the models to determine which predictions were most sensitive to the amount of training data and what taxa or taxa pairs were most important for model prediction. The authors also incorporated extra experiments after learning on the original data to then test how well their model could predict functional capacity on new test data. The main findings were that Bacteroides has broad metabolic capability with the model highlighting specific species with more specialized metabolic capabilities.

Strengths:<br /> - Paper integrates extensive experimental data.<br /> - Sensitivity analysis was performed on the model (which is often neglected), the reviewer appreciates this extra step.<br /> - LSTM model was in python and notebooks could be downloaded and run with ease.

Points of weakness:<br /> - It is unclear why an LSTM would be a good model for the microbiome<br /> - It is unclear what aspects of the dynamics are long-term, and whether the experiments capture this long-term effect<br /> - Discussions around the LSTM model and some ML and dynamical systems concepts are inaccurate (LSTM with one hidden unit is not really a "deep" model, gLV models are linear in the parameters and thus the parameters are trivial to solve for give the microbial abundances)<br /> - Not enough detail is given regarding the LSTM model or the composite model to understand them<br /> - part of the composite model is in Matlab and could not be tested<br /> - authors claim that their model is interpretable, but it is no more interpretable than any differentiable model that can use gradients to open the lid after training

An appraisal of whether the authors achieved their aims, and whether the results support their conclusions: The authors are commended on their extensive experimental integration and some aspects of validation. The models however are missing enough details in the text to understand how they were used. Also, the comparison seems a little unfair. From reading the text it appears that the LSTM+FF model was trained jointly, whereas, the composite model first learns from the microbiome data and then the metabolite prediction component is trained after the gLV model parameters are held fixed. Any model trained jointly will have an advantage to one trained in this two-step process. If the main claim of the paper is that the LSTM model is better than a gLV model then the comparison should be more systematic and fair.

Likely impact on the field: The tight coupling of new experiments with computational methods is important. All too often a tool is made but only shown to work on data not tailored to the tool. Here, both are designed together.

Reviewer #1 (Public Review):

This manuscript aimed to develop the animal models for Pitt-Hopkins syndrome (PHS) by showing a number of behavioral data and electrophysiological recordings. The major strength is employing sophisticated transgenic Tcf4 mouse model and high-quality of data. The limitation is the marginal rescue effects by Tcf4 reinstatement and murky interpretation of some results. Given the widespread expression of Tcf4 in central nervous system, extreme caution is required for considering the current mouse model for PHS pathogenesis.

Reviewer #2 (Public Review):

Mutations in the TCF4 gene leads to Pitt Hopkins Syndrome, a neurodevelopmental disorder. This study demonstrates that restoring the typical TCF4 gene in most, but not all, cells of the brains in a mouse model of TCF4 gene deletion can have a beneficial effect on the behavioral symptoms of Pitt Hopkins Syndrome if given very early in brain development. Much more work is needed to determine if such an approach is actually feasible in patients because current gene therapies are not able to be performed in the same manner in humans as was done in this mouse study. Additional future studies will be needed to determine if such a treatment approach in mouse models will be useful later in brain development. Additional controls for effects of the virus used for their mouse gene therapy would have been useful.

Reviewer #1 (Public Review): 

In this study, the authors took an experimental, empirical approach to tackle the thorny problem of micro-scale variation in soil properties within and among field plots in confounding statistical analyses. The issue is that in field experiments, small variation in one or more soil property variables can obscure true effects of experimental variables on plant phenotypes. In this case, the authors used Sorghum (many accessions) as the focal plant and a drought vs. well-watered treatment as the main experimental variable. Really, this means genotype x drought treatment was the central question of the study. Many other variables were measured, including the microbiome (the authors cite a preprint where this part of the study is explained in more detail, although unless I am confused, it seems as though there was an experimental treatment of microbes in the biorxiv study--was this also the case in the present study or were the microbes naturally colonizing the rhizosphere?). Overall, the PC-based approach to de-noise these kinds of datasets is sound and provides an important advance in the sense that pulling out subtle phenotypic effects in field trials may now be more straightforward given the results of their study and the tools that they provide. The main result is that without their framework they would not have found the association between water treatment, plant growth and Microvirga bacterial abundance--it would have been lost to the noise inherent in these kind of large-scale experiments with relatively modest degrees of freedom.

Reviewer #2 (Public Review): 

The authors present a simple statistical workflow to strengthen biological signals of interest by accounting for spatially-structured environmental heterogeneity in field settings. Oftentimes, environmental variation is not neatly partitioned among sections of an experimental plot (e.g., rows or blocks). As a result, statistical models that do not account for the shape of environmental "noise" across the landscape will poorly capture (and poorly control for) its confounding effects. 

The presented approach addresses this problem by testing for spatial structure in each of a range of assayed environmental variables, collectively capturing the numerous spatially structured variables into a few principal components, and regressing out their effects on experimental outcomes. This is noteworthy because it focuses on measuring a suite of environmental characteristics and modeling their collective effects on the outcome directly, rather than attempting to account for them indirectly by modeling spatial variation in the outcome itself. The authors justifiably posit that their approach, in comparison to spatially-unaware models, has the potential to (1) boost signals of interest by reducing background noise and (2) reduce false positives that arise when unmodeled environmental variation correlates with the spatial distribution of treatment effects. Their application of this approach to a rich empirical dataset offers an opportunity to explore its utility. 

It should be noted that no major component of this approach is new, even in the very specific case of soil elemental composition and plant field trials. For example, Pauli et al. 2018 (G3, doi: 10.1534/g3.117.300479) used very similar methods to measure soil elements, interpolate missing data points, and account for the local soil environmental effects on phenotypes of interest. Additionally, Murren et al. 2020 (Am. J. Bot., doi: 10.1002/ajb2.1420) used principal components regression to reduce the dimensionality of environmental variation (including soil elemental properties) and quantify its effect on plant traits (in this case, fitness). The main advance of the current paper is demonstrating that integrating these approaches is a simple and effective way to address environmental variation that poses a nuisance to the study. 

Strengths:  

The authors describe a potential way to boost power and reduce false positives -- without the costly (and yet still imperfect!) approach of massively increasing experimental sample sizes. Additionally, the authors demonstrated that environmental states at un-measured locations can be successfully interpolated from spatial relationships among the sampled locations, so their approach can be used even when it's infeasible to measure environmental variables for the majority of samples. Finally, their use of dimensionality reduction for environmental features (in this case, principal components analysis) allows very rich environmental profiles with many variables to be included in an analysis, without greatly increasing the complexity of the statistical models used to test for treatment effects. Overall, these advantages make the approach feasible and broadly applicable across field studies. 

Another advantage of this approach, relative to methods that account for spatial variation in outcomes without modeling environmental contributors, is the potential to reveal mechanistic insights. For example, the authors identify specific soil properties (e.g., phosphate levels) that correlate with individual plant or microbial community traits. On one hand, this naturally generates additional hypotheses to test through future experiments. On the other hand, it allows researchers to leverage previously-known mechanistic insights into their experimental systems when choosing which environmental features to measure. 

The authors demonstrate how their approach can be applied to a rich empirical dataset. As dependent variables, the dataset includes plant harvest traits, leaf traits, metabolomic profiles, and microbiomes of the plant root, adjacent soil, and root-soil interface. As independent variables, it includes plant genotypes, drought treatments, and their interaction. And as environmental variables, it includes soil composition properties. The authors demonstrate that a subset of the soil composition properties are spatially structured, and that accounting for their effects yields new insight. For example, the authors identify an OTU that is correlated with increased plant height, suggesting potential growth-promoting effects, but only when adjusting both OTU abundance and plant height for soil properties. Such promising results suggest that modest effort to measure and model spatially structured environmental variation can illuminate findings in real experimental settings that would otherwise be obscured, and that integrative approaches to measure a variety of environmental features can add significant value to a study. 

Weaknesses:  

The paper lacks some components that I would expect to see when presenting an analytical approach and arguing for its effectiveness. Because the study's purpose is framed as providing "a tool with which sources of environmental variation in field data can be identified and removed, improving our ability to resolve effects of interest and to quantify subtle phenotypes", critically and thoroughly evaluating its performance is of utmost importance. Below, I describe a few reasons to be a bit cautious when interpreting the results. 

First, the extent to which the approach improves inferences of treatment effects is not comprehensively shown. Rather, aspects of a few of the most promising results from a series of tests are presented. A presentation of the method's performance across the full series of tests conducted (i.e., for each dependent variable considered for a given model formulation) is needed to truly understand how it strengthens the analyses. This would include a comparison of each statistical model before and after adjusting for environmental variability. Metrics such as the increase in variance explained by experimental treatments and the increase in the proportion of treatment effects that are significant, across the full range of tests, would complement the current presentation and better demonstrate the extent to which the approach boosts signals and reduces unexplained noise in field data. 

Second, comparisons to other approaches to account for spatial variation in field trials were not conducted, and a strong conceptual discussion of how, when, and why its performance may differ from these other approaches is lacking. Some other methods (e.g., Velazco et al. 2017, Theor. Appl. Genet; doi: 10.1007/s00122-017-2894-4) model how each focal observation differs from expectations based on spatially neighboring observations, but do not measure or directly model the sources of environmental variation (e.g., soil properties). This may have different strengths and weaknesses. For example, if the environmental variables with the largest effects are not measured, they would be missed in the current study's approach but could be indirectly accounted for by other methods. Knowledge about the relative performance of different approaches under realistic scenarios (perhaps both empirical and simulated) would be important to researchers considering use of the presented approach, who must gauge if the additional effort and resources needed to measure environmental parameters are likely to improve statistical inferences beyond other spatially-aware statistical methods that do not require measuring such covariates. 

Thus, while promising results in support of this approach are presented, a reliable picture of its overall effectiveness will require further investigation.

Reviewer #3 (Public Review): 

This work provides a rigorous strategy to reduce the noise in data collected from field studies. Based on a sorghum field trial, the authors proposed a tool to identify and estimate the effects of soil properties with minimal degrees of freedom. Such a procedure can help us better understand the phenotypes of interest using appropriate statistical analyses. 

Strengths: Accounting for the confounding effects from all kinds of variables is crucial for the analysis of field data. The tool presented in the manuscript can be used for field studies in general and furthermore, extended to other types of trials with minimal modifications. 

Weakness: It will be helpful to include the details of the statistical model in the manuscript.

Reviewer #1 (Public Review): 

Kozhemiako et al. characterized several NREM sleep parameters, including their relationship with each other and with waking event-related potentials and symptom severity in patients with schizophrenia (SCZ) relative to healthy control (HC) subjects. The authors confirmed a marked reduction in sleep spindle density in SCZ while also showing that only slow spindles predicted symptom severity, and that fast and slow spindle properties were largely uncorrelated. Also, the main sleep findings were replicated in a different sample, and a model based on multiple NREM components predicted disease status with good accuracy in the replication cohort. Furthermore, despite being altered in patients with SCZ relative to HC, auditory event-related potentials elicited during wakefulness were unrelated to NREM sleep parameters. Based on these findings, the authors concluded that the present study lays the foundations for assessing these sleep and wakefulness EEG neurophysiological markers, individually or in combination, to guide efforts at identifying individuals with SCZ, and especially those who are most likely to benefit from specific treatment interventions. 

This study has several strengths, but certain aspects of the data analyses and of the interpretation of the main findings need to be clarified and extended. 

Strengths:

The authors conducted the largest replication study of sleep findings in patients with SCZ relative to HC. One of the main challenges in clinical research nowadays is confirming previously established findings (i.e., reduced sleep spindle density in patients with SCZ) in a different group of patients. The authors should therefore be commended for doing so in a quite large sample of SCZ patients. It should also be pointed out that they were able to replicate most of the sleep findings in a demographically distinct sample of patients with SCZ. 

Another strength of the present study is the assessment of novel sleep spindle and slow oscillation (SO) measures in patients with SCZ relative to HC. In addition to a comprehensive characterization of previously known spindle and SO parameters, here the authors introduced some novel measures, including the intra-spindle frequency modulation (chirp/deceleration) and its relationship with the SO phase as well as the Phase Slope Index (PSI) as an index of functional connectivity. 

The assessment of wake EEG abnormalities in the same group of SCZ patients showing altered sleep parameters is another strength and novelty of the present study. Neurophysiological alterations during wakefulness, assessed with Mismatch Negativity (MMN), auditory P50 S2/S1, and auditory steady state response (ASSR) power have been previously reported in patients with SCZ. By computing these wake neurophysiological parameters along with sleep EEG measures in this study, the authors were able to investigate whether these wake and sleep abnormalities were associated with or rather reflected distinct alterations in underlying neural circuits in SCZ. 

Finally, by using a Joint model analysis across sleep EEG metrics that were altered in SCZ, the authors were able to establish an excellent ability in predicting case/control (SCZ/HC) status in the training sample along with a good ability in predicting SCZ/HC status in the independent target sample. 

Weaknesses:

An important finding of this study was the correlation between sleep spindles and severity of symptoms. The authors should, however, report whether this correlation between slow spindles and clinical symptoms was confirmed in the replication sample of SCZ patients. 

The authors computed novel sleep measures, some of which were altered in patients with SCZ relative to HC. For example, a decrease in overlap between slow spindles and SO (a proxy measure of spindle=SO coupling) as well as an increase in PSI (a proxy measure of connectivity) was reported in SCZ patients. However, the relevance and the functional implications of these alterations are barely addressed in the discussion. 

In the abstract, as well as towards the end of the discussion, the authors suggest that the present findings may index risk, sequelae, or modifiable therapeutic targets. Each of these claims needs to be further elaborated on.

Reviewer #2 (Public Review): 

This study sets out to replicate the large and accumulating literature which shows alterations in sleep neurophysiology in individuals with schizophrenia. The strengths of this study include the sample size and the analyses performed, which are thorough. 

One limitation of the work is that too many analyses are presented which do not contribute to the overall "story" of the paper. It has long been known that different features of sleep, such as slow oscillations and spindles, map onto somewhat distinct networks and that additional information can be derived from combining these measures. Therefore, the section on PSC analysis can be reduced. Furthermore, the value of a "replication" sample is not clear. These previously published data have already shown spindle deficits in their samples, so the argument is rather circular here. The additional analyses which were done with the "replication" sample do not add significantly to our knowledge of the neurobiology of schizophrenia.

Reviewer #3 (Public Review): 

Understanding of the mechanisms of sleep alterations in patients with schizophrenia, may provide important information for the development of new therapies for psychosis. The main strength of this study is that it provides a most comprehensive analysis of sleep EEG in patients with schizophrenia. The results presented are generally consistent with the existing knowledge. The main weakness of this study is that it fails to take into account the potential contribution of sleep homeostasis and circadian rhythms, as well as relevant environmental factors, such as light.

Reviewer #1 (Public Review):

This is an interesting and scientifically sound study, for the first time demonstrating the important role of Cdh5 expression in trophoblasts using a genetic model. The wider interest in this study relates to the potential impact of VE-cadherin's guidance role in trophoblast invasion, for preeclampsia pathogenesis.

VE-cadherin (gene name, Cdh5) is well known for its essential role in endothelial cell-cell contacts. However, placenta trophoblasts also express Cdh5. Sung et al., explore the effect of deleting Cdh5 from trophoblasts using a Cyp19-Cre mouse model. Loss of trophoblast Cdh5 results in suppressed invasion of Cdh5- trophoblast into spiral arteries, loss of spiral artery remodeling, reduced maternal blood flow into the placenta and disturbed embryonic development. While there is no effect of Cyp19-Cre-dependent Cdh5 deletion on endothelial Cdh5 expression or blood vessel density, trophoblast density and invasion depth are markedly suppressed. As a result, embryonic development is severely disturbed. The results are appropriately controlled and the authors' conclusions are well underbuilt.

Reviewer #2 (Public Review):

Fetal trophoblasts are known to invade the mammalian decidua in the placenta and remodel and connect with maternal spiral arteries. This paper shows that VE-cadherin on trophoblasts is needed for the invasion of trophoblasts into the decidua and for the remodeling of spiral arteries. It was shown before, that fetal trophoblasts express endothelial adhesion molecules (such as PECAM-1 and VE-cadherin) and it was shown in in vitro migration assays that blocking or genetic inactivation of VE-cadherin interfered with trophoblast invasion/migration. The current work shows that Cyp19 (cytochrome P450)-Cre driven genetic inactivation of VE-cadherin in trophoblast cells interferes with embryonic development, invasion of fetal trophoblast into the decidua, supply of the placental labyrinth region with maternal erythrocytes, smooth muscle cell association of spiral arteries (SA) and endothelial cell displacement in SA by trophoblast cells. In addition, Doppler ultrasound analysis revealed that Cyp19-Cre driven VE-cadherin gene inactivation results in placental insufficiency and fetal distress.

The study is important since it shows for the first time a causal link for the need of VE-cadherin on trophoblasts in vivo for the invasion of these cells into the decidua and for their role in vascular remodeling. The conclusions of this paper are mostly well supported by data, but some aspects of this study need to be extended.

1) It is not entirely clear whether the expression of Cre driven by a very small 501 bp promotor fragment of the Cyp19 gene is indeed sufficiently specific for trophoblasts in the placenta. The paper describing the generation of these Cyp19Cre mice mentions that the Cyp19 gene which encodes cytochrome P450 is expressed in various organs, among them the fetal liver. Hemorrhaging found in the E12.5 embryo could be an indication for endothelial loss of VE-cadherin in Cyp19Cre;Cdh5fl/fl embryos. Although suppl. figure 1 shows some controls for VE-cadherin staining of endothelium in the placenta of these mice, it is unclear whether this holds up for other organs.

2) Immunofluorescence staining of the paraffin sections resulted in high quality images in figures 2 and 3, which are very instructive. However, it is difficult to display a representative overview of the whole tissue if the results are simply based on thin sections. The results could vary a lot between sections at different positions. Therefore, the study would benefit from whole mount staining of either thick vibratome sections or even larger tissue parts that went through a tissue clearing procedure. This would allow getting a more representative 3D picture of the vasculature in large parts of the placenta. In addition to providing a more representative picture of the tissue, this would also allow us to see how the trophoblast cell system connects to spiral arteries beyond single cells integrating into the endothelial cell layer of the vessel wall. While the paraffin sections indeed very nicely show how trophoblast cells integrate into spiral arteries, which looks like it could be the result of "displacement" of endothelial cells, it is unclear how this would link to a fetal trophoblast/vascular system. This question can only be answered by analyzing whole mounts.

3) The paper does not provide any mechanistic studies that could explain how VE-cadherin on trophoblasts could support the invasion of trophoblasts into the decidua and how it could be responsible for the dissociation of smooth muscle cells from spiral arteries. It was shown before that silencing of VE-cadherin in a trophoblast cell line interfered with the invasion/migration of these cells in in vitro assays. It would be great if the present study would attempt to analyze the consequences of VE-cadherin silencing in such cells on the expression pattern of migration relevant genes in trophoblastic cells. Is collective cell migration affected or is it a defect in cell dissociation that impairs trophoblast dissemination into the decidua of the placenta?

Reviewer #3 (Public Review):

During pregnancy, fetal trophoblasts invade the maternal decidua and remodel spiral arteries to bring maternal blood into the placenta. This process, known as endovascular invasion, is thought to involve the adoption of functional characteristics of vascular endothelial cells (ECs) by trophoblasts through a process termed vascular mimicry. Invasive trophoblasts in preeclamptic placentas lack VE-cadherin, and loss of VE-cadherin reduces trophoblast invasion in vitro, suggesting the critical roles for VE-cadherin in endovascular invasion and vessel formation. Based on the findings, the authors claim a non-endothelial role for VE-cadherin in trophoblasts during placental development and suggest that endothelial proteins may play functionally unique roles in trophoblasts that do not simply mimic those in ECs. Overall, the data are analyzed thoroughly and the conclusions drawn are novel and appealing. Understanding molecular and cellular pathways for endovascular invasion and pathogenesis of preeclampsia are important topics for current vascular biology and Ob/Gyn biology, making this study timely and important. It would be constructive if the authors provide the underlying molecular mechanism.

Reviewer #1 (Public Review):

This manuscript addresses the role of HH signaling in larynx development by combining mutant analysis and RNA-sequencing. The authors show that the distinct domain of low Shh expression and induction of N-Cadherin expression at epithelial lamina and esophageal epithelium of the larynx during larynx-esophageal separation. The authors then show that loss of E-Cadherin and upregulation of N-Cadherin accompanied by differentially expressed EMT-related genes in larynx tissues of Shh null embryos, suggesting that Shh regulates larynx morphogenesis through inhibiting cadherin switch and EMT. Notably, Shh-descendant cells undergo transition to a mesenchymal fate in Shh mutant larynx. Additionally, Shh mutant embryos display loss of Nkx2.1 and reduced Sox2 expression in larynx epithelial cells at early laryngeal morphogenesis and disruption of larynx-esophageal separation accompanied by disorganized and thickened epithelium at later stages. The authors find ectopic Pax1 expression in the larynx of Shh mutant, suggesting expansion of pharyngeal pouches in the absence of Shh. Taken together, these characterization studies provide interesting findings and important information for the role of HH signaling in the development of the anterior foregut.

Reviewer #2 (Public Review):

The authors present genetic and cellular evidence for a role for Shh in repressing EMT, such that in the absence of Shh cells undergo EMT and are replaced with a novel epithelial cell population of unknown origin. The methodology is straightforward and although the paper is essentially a standard mutant phenotype analysis, given the relative lack of knowledge about how this important structure develops these results make a potentially significant contribution to field.

The first part of the study investigates how Shh signaling normally is deployed during larynx development, using an allele of Shh that expresses both GFP and Cre, allowing both real-time gene expression reporter and lineage trace functions in the presence of normal SHH signaling.They then investigate whether complete loss of Shh impacts these same phenotypes. They show using IHC and RNA-seq the upregulation broadly of EMT-associated markers, and the localized cadherin switch and vimentin expression consistent with EMT, as well as the morphological appearance of what appear to be extruded cells within the lumen. They then document the ontogeny of this cadherin switch, and confirm the origin of these cells using the ShhCre lineage trace, as well as demonstrating they undergo apoptosis at high rates after leaving the epithelium. These cells also downregulate Foxa2, a transcription factor known to suppress EMT. These data are all straightforward, quantified, and reasonably interpreted.

The more interesting aspect of this mutant analysis is their investigation of why this increase in EMT and apoptosis doesn't result in loss of epithelial integrity. They revisit Figure 4 and point out that only the ventral pharynx is Shh lineage-positive at E10.5, even though it is initially expressed throughout the pharynx. Their conclusion that this result implies "that foregut epithelial cells undergo dynamic regional cellular rearrangements during this timepoint" is a bit vague, and it isn't actually clear what they mean - what kind of rearrangements, and where do these non-Shh lineage cells come from? They then show that there are very few Shh lineage positive cells in the mutants at this stage (although the quantification isn't quite correct - 80% positive to 20% positive is a loss of 75% of the Shh lineage cells - and the image shown only has a couple of TdT+ cells, not consistent with the 20% their quantification in panel L shows).

The authors then explore where these Shh lineage-negative cells come from. The authors start by suggesting that perhaps they are dorsal-identity cells that spread, and use Sox2 and Nkx2.1 as dorsal and ventral identity markers. However, both of these markers are reduced in the mutants, leading them to conclude that Shh is required for regional identity of both dorsal and ventral cells. However, this doesn't answer the question of whether dorsal cells spread ventrally, and the shape of the pharynx and the epithelium itself are both really abnormal, so it's just unclear what is happening here.

They then look at RNA-seq data again for a clue and find that these cells have up-regulated pharynx-related genes most of all, the highest of which is Pax1, normally a pharyngeal pouch-specific transcription factor. They propose several possibilities for the origin of these cells, but do not address this question. As written, this is an interesting observation but not really explored.

Reviewer #3 (Public Review):

The early foregut epithelial cells express Shh, which plays a crucial function in maintaining epithelial integrity. The authors observed that Shh expression is dynamic and speculated that spatio-temporal control over its expression might regulate regionalized remodeling. The domains undergoing remodeling overlap with reduced Shh expression in wild-type tissue. Additionally, these domains were marked by defects associated with disruption of epithelial integrity along with an upregulation in EMT (Cadherin switching and RAB11) markers. To support their findings, the authors demonstrate that Shh-/- epithelial cells undergo premature/ aberrant EMT which also leads to disruption of epithelial integrity. The null mutants showed disruption of basement membrane, extrusion of epithelial cells into the mesenchyme and defects in cell survival. Last, they document that the Shh mutant epithelial cells are replaced by a new population of cells, but while these cells may be of pharyngeal origin this is not demonstrated.

Reviewer #1 (Public Review):

This is a very interesting and potentially very important paper that shows multi-organ effects of JUUL exposure in mice. I have some major and minor comments which are listed below.

Major Comments:

I am concerned that a lot of these studies had relatively low n numbers (n=5 in some cases) and that some of the studies may have been underpowered. Given the variability with in vivo studies, some endpoints may have been significant with more numbers. Along these lines, what is the justification for using the (parametric) ANOVA test. I'm not a statistician but I thought that the rule of thumb was that non-parametric tests should be used if n<12 since you cannot verify that the data is normally distributed. In this case, I would recommend having a statistician look at it and/or increasing some of the N's, or using the non-parametric Kruskal-Wallis test. Indeed, in some cases, the variation the variation is quite large (i.e. Fig 6, 7). Whilst I do not think that the low N's change the ultimate conclusions, but more rigor (i.e. more N's) would help solidify the paper given that it will likely be of great interest and scrutinized by the scientific community.

Fig S3. For the lung histology, please quantify the mean linear intercept per ATS guidelines and show representative BAL images.

One of the must novel conclusions from this paper is increased inflammation in the brain which the authors speculate could lead to altered moods and or change the addiction threshold. I would tend to agree with this conclusion, but could the authors perform additional mouse psychological tests to confirm this? Also, were there observable physiological responses in the vaped mice that could be reported which may correlate this conclusion, i.e. changes in grooming, fur ruffling or other behavioural changes?

Minor comments:

Change title to state "in mouse". That this study was performed in rodents should be apparent from the outset.

No changes in collagen deposition were detected using basic histology. Have the reviewers considered performing immunohistochemistry and staining for alpha-smooth muscle actin which may be a more sensitive assay?

"Thus long term exposure to Juul does not lead to significant changes...". I would argue that 1-3 months is not long term. Indeed, other researchers have performed 6-12 month e-cigarette exposures and it takes a lifetime in humans to develop lung disease after smoking. Since you can detect pro-inflammatory changes but no altered physiology, it may be that alterations in airway physiology are only just beginning.... The authors should modify this sentence and maybe not call their studies "long term".

"Differences in LPS induced cytokine levels were no longer observed after 3 month JUUL exposure versus Air control groups". As per the major comments, this might be a power issue - there is certainly a trend for some cytokines.

Reviewer #3 (Public Review):

In this study, Alex Moshensky et al. investigated effects of chronic aerosol inhalation of flavored JUUL on inflammatory markers in several organs, including brain, lung, heart, and colon in a mouse model. They found that JUUL inhalation upregulated a number of cytokine and chemokine gene expression and increased HMGB1 and RAGE in the nucleus accumbens. Inflammatory gene expression increased in colon, and cardiopulmonary inflammatory responses to acute lung injury with LPS were exacerbated in the heart. They also found flavor-dependent changes in several responses.<br /> Overall, it is a descriptive study and the conclusions was not clearly supported by the data.

Strengths:

Due to the rapid evolution of vaping devices, the data on health effects of Pod devices are scarce. This study provides useful information on the inflammatory change caused by chronic JUUL aerosol inhalation.

Weaknesses:

1. The authors observed neuroinflammation in brain regions responsible for behavior modification, drug reward and formation of anxious or depressive behaviors after exposure to JUUL. The importance of the neuroinflammation is still unclear. It would help demonstrate the pathogenic role of the neuroinflammation by testing animal behaviors. Similar issue for other organ inflammation.

2. Majority of the data are inflammatory cytokine mRNA expression. Other methods would be needed to confirm their expression.

3. The author seemed to assume the difference between JUUL Mango and JUUL Mint is flavor and then came up with the conclusion regarding flavor-dependent changes in several inflammatory responses. Evidence is needed to approve the assumption.

4. In most cases, the change of inflammatory cytokines is mild ~2 fold. The author should demonstrate how these marginal changes could affect pathophysiology.

5. To fully evaluate the health impact of evolving cigarette, it would be informative to included other tobacco or vaping device as control.

6. The longest exposure in the study is 3 months. It is not convicting to come up with conclusions regarding chronic exposure. Some organ showing no difference may be due to the timing.

Reviewer #2 (Public Review):

Under homeostasis conditions, the authors observed sign of inflammatory responses in the brain, the heart and the colon, while no inflammation was detected in the broncho-alveolar lavage fluid of the mice following exposures to JUUL aerosols. Also, JUUL aerosol exposures mediated airway inflammatory responses in the acute lung injury model (LPS). Further, this infection affected the inflammatory responses in the cardiac tissue. Most of the biological adverse effects induced by JUUL aerosols were flavor-specific.

Strengths include evaluating inflammation in multiple organs, as well as assessing the physiological responses in the lungs (lung function) and cardiovascular system (heart rate, blood pressure), following exposures to JUUL aerosols. Weaknesses include the fact that only female mice were used in this study. Further, the daily exposures to either air or to the JUUL aerosols lasted only 20 min per day. It is unclear how a 20-min exposure is representative of human vaping product use. Also, although daily exposures were conducted for a duration of both 1 and 3 months, time-course effects associated with JUUL aerosols are barely addressed.

Although there are a few limitations related to this study, which should be included in the manuscript, overall, the authors' claims and conclusions are based on the data that is presented through multiple figures.

Reviewer #1 (Public Review):

This manuscript describes a clever strategy to nominate existing drugs for testing for their ability to abrogate the well-known negative effects of platinum-containing chemotherapeutic drugs on hearing. This is an important clinical problem, and as such if existing drugs actually had the ability to mitigate hearing loss when given in combination with cisplatin (assuming the combination is well-tolerated and the combination does not mitigate the anti-cancer effects of cisplatin), the potential for clinical impact would be substantial.

The authors first define a gene expression signature of cisplatin resistance, using publicly available gene expression datasets of cell lines that are either sensitive (meaning are killed by) to cisplatin, or are resistant to it. They then use that signature of resistance to query the Connectivity Map, which is a publicly accessible database of gene expression profiles following genetic and small molecule perturbation. The authors thus aimed to find drugs that phenocopy the cisplatin-resistance signature.

This analysis yielded a large number of candidate drugs, which was whittled down to a smaller number (30) based on a number of bioinformatic analyses. These top candidates were then tested in cellular and animal (zebrafish and mouse) models of cisplatin-induced hearing loss, and found that the vast majority (87%) showed some evidence of hearing protection in at least one of the in vivo models, and half of the top compounds scored in both in vivo assays. The FDA-approved drug niclosamide was explored in greater detail.

The strengths of the study include a) its tackling of a clinically important problem, b) its clever use of the Connectivity Map for discovery, and c) the validation of top compounds in what appear to be relevant in vivo models of cisplatin-induced hearing loss.

The weaknesses of the study include a) the lack of any mechanistic insights into how niclosamide and other top scoring compounds are acting, and b) the lack of a thorough testing of niclosamide's lack of an ability to cause cisplatin resistance with respect to cisplatin's anti-cancer activity.

Reviewer #2 (Public Review):

This study performed in silico screens on Connectivity Map (CMap) to identify transcriptomic profiles of cisplatin-resistant and cisplatin-sensitive cancer cell lines for the purpose of identifying and repurposing FDA-approved compounds as potential otoprotectants against cisplatin-induced hearing loss. Niclosamide, an FDA-approved drug for tapeworm treatment with a favorable safety profile, was identified through this screening method and validated using mice and zebrafish, demonstrating protective effects against cisplatin- and noise-induced hair cell loss and partial protection from hearing loss.

The premise of using transcriptomic profiles of cancer cell lines to model the profiles of inner ear hair cells in response to FDA-approved compounds has not previously been explored and provides a novel screening method to identify potential otoprotective therapies. Yet this reviewer is not entirely convinced the rationale for selecting niclosamide exclusively for further investigation was well reasoned. Additionally, the expansion of this study to further investigate the ability of niclosamide to protect against noise-induced hearing loss seemed far removed from the initial intent of the in silico screens. A more focused and rigorous evaluation of several candidate compounds and their ability to protect against cisplatin ototoxicity would greatly strengthen this study.

Reviewer #1 (Public Review):

This study confirmed the role of BCL6 in tumor cells escaping cytotoxic drugs in vitro. The upstream and downstream molecules of BCL6 were determined by in vitro experiments. The role of BCL6 inhibitor in chemo-sensitization was verified by in vitro experiments and animal experiments，which proved certain clinical application value. These findings establish a rationale for targeting BCL6 to conquer resistance to genotoxic stress in solid tumors which has certain clinical value. However, ample researches have reported during the past years, innovation of this manuscript is required to be addressed. The authors have done a lot of research and obtained abundant data which provides strong support for the research conclusion.

Reviewer #2 (Public Review):

In the paper entitled "The Oncoprotein BCL6 Enables Cancer Cells to Evade Genotoxic Stress", through comparing transcriptional profilings of ETO sensitive versus resistant tumor cell lines, the authors found that BCL6 was selectively upregulated in ETO-resistant tumor cells, and their further in vitro and in vivo data suggest that Bcl6 upregulation via the IFN-STAT1-Bcl6 axis conferred tumor resistance to genotoxic stress, and targeting Bcl6 significantly improved therapeutic efficacy of ETO/ADR in mouse tumor models.

Their findings are interesting and may inspire new combinational therapeutic strategy in treating chemotherapy resistant cancers, although a number of issues remain to be further clarified.

Major concerns:

1. Through using in vitro assays, the authors defined a panel of genotoxic agents (ETO, ADR, etc) resistant or sensitive tumor cell lines, and indicated the resistance was caused by BCL6 upregulation. It was expected in the following on animal studies, the authors would choose tumor cell lines with clearly defined phenotypes characterized in their study. But it was not the cases in their studies. For examples, in Fig S2C and Fig 7B, the authors used an ambiguous tumor cell line HCT116 to test ETO resistance, which had only a borderline level of resistance to ETO (Fig 1A) but yet sensitive to ADR (Fig S1A), whereas in Fig 2H, the authors chose a tumor cell line (MCF7) not examined in their study, instead of the high ETO-resistant tumor cell lines H661/Capan-2 or high ADR-resistant cell lines DLD-1/H836.

2. Fig 3, the concept of tumor cell expressing IFNa/IFNg conferring genotoxic resistance sounds very interesting and novel, but the authors only tested IFNa/g expression at transcriptional level, protein expression data should be also provided.

3. Fig 3F-3I, ETO-induced interferon response should be examined comprehensively in different tumor cell lines as listed in Fig 1A/2A. Similarly, effect of exogenous IFNa/IFNg on ETO-resistance should be also examined comprehensively in both sensitive or resistant tumor cell lines. In addition, the effect of blocking IFNg/IFNa on ETO-resistance should be also tested in different tumor cell lines. These data are extremely useful for extending or strengthening the broad impact or influence of their findings.

4. Fig 4A-L, the authors examined activation of IFN-STAT1-Bcl6 axis in tumor cells in different angles via different approaches, but using different tumor cell lines in different panels of experiments, making it quite annoying and difficult to judge their findings across different tumor cell lines. At least, ETO or IFNa/IFNg induced STAT1 upregulation and its phosphorylation should be examined comprehensively in both resistant and sensitive tumor cell lines.

Reviewer #3 (Public Review):

Overcoming intrinsic and acquired drug resistance is a major challenge in treating cancer patients because chemoresistance causes recurrence, cancer dissemination and death. In this manuscript, Liu et al. reveals oncoprotein B cell lymphoma 6 (BCL6) as a core component that confers tumor adaptive resistance to genotoxic stress. They show that transactivation of BCL6 by genotoxic agents is associated with a poorer therapeutic efficacy and clinical outcome. Genotoxic agents lead to the transcriptional reprogramming of pro-inflammatory cytokines, which subsequently upregulates BCL6 expression. Then, BCL6 represses PTEN and consequently leads to drug resistant. Accordingly, the BCL6 inhibition enhances etoposide-triggered DNA damage and apoptosis. The conclusions of this paper are mostly well supported by data, but some aspects of image acquisition and data analysis need to be clarified and extended.

Reviewer #1 (Public Review):

DDX39B (also BAT1 and UAP56) has previously been shown to be a risk factor in Multiple Sclerosis, as well as other autoimmune diseases, and as a regulator of IL7 splicing. These data led the authors to postulate a broader role of DDX39B in immune regulation. They first carried out RNA-Seq from primary human CD4+ T cells depleted of DDX39B and found ~700 genes that are differentially expressed in a DDX39B-dependent manner, including about 10% of known MS-susceptibility genes detected. One of the most robustly identified DDX39B-sensitive genes in this experiment was the gene encoding the Fox3p transcription factor, a result that was reproduced in primary and cultured Treg cells. Consistently, loss of Fox3p target genes was also observed.

The authors go on to use RNA-Seq, subcellular fractionation and shRNA-resistant rescue experiments, conclude that reduced Fox3p expression in DDX39B-depleted cells is due to inefficient splicing of introns with C-rich polypyrimidine tracts. This is consistent with prior work suggesting a role of DDX39B in 3' splice site selection, although prior studies have not defined what specific role this protein plays. More broadly, the authors compare DDX39B-sensitive and resistant introns transcriptome-wide, and also find evidence for increased C-richness within DDX39B-dependent 3' splice sites - although they are careful to emphasize this sequence bias is not sufficient to confer DDX39B dependency.

Reviewer #2 (Public Review):

In this manuscript, the authors have implicated the RNA helicase DDX39B, a gene associated with increased MS susceptibility, in Foxp3 splicing. In particular, the authors demonstrated that DDX39B is required to reduce Foxp3 intron retention, thereby promoting its expression, which taken together is expected to increase Treg functionality in the context of Multiple Sclerosis. The authors have used human total CD4 cells, the Treg-like cell line MT-2 and in some context, primary human Treg cells to perform these experiments. The biochemical experiments in my opinion are well done but primarily done in cell lines which include MT-2 as well as HEK293 and HeLa cells. Understandably, it is difficult to do these experiments in primary T cells, let alone Treg cells. However, at least some experiments with in vitro generated iTreg cells, along with functional experiments, would greatly strengthen the work.

Reviewer #1 (Public Review): 

In this study, Mead and colleagues report that global loss of ADAMTS6 causes a severe chondrodysplasia that is significantly worsened by concomitant loss of ADAMTS10 and, conversely, almost fully prevented by haploinsufficiency for fibrillin 2, a substrate of ADAMTS6. Of note, haploinsufficiency for fibrillin 1 does not affect the chondrodysplasia of ADAMTS6 null mice. The authors use a variety of in vivo and in vitro assays for the testing of their hypothesis. 

The paper is informative as it expands and deepens our current undertstanding of proteases and their substrates in endochondral bone development. 

Though the phenotype is interesting and the rescue experiment is compelling, it remains completely elusive how the loss of ADAMTS6 and the increased accumulation of fibrillin 2 cause severe chondrodysplasia, which negatively impinges on the novelty of the paper. 

In additon, numerous technical issues should be addressed to strengthen the authors' conclusions and their biological relevance.

Reviewer #2 (Public Review): 

This work investigated the role of Adamts6, a protease closely related to Adamts10, in fibrillin proteolysis and demonstrates that Fbn2 digestion by Adamts10 and Adamts6 plays a critical role in skeletal development. The study shows the overlapping roles of these proteinases in fibrillin digestion and elimination and in skeletal development. Using mouse genetic approaches, the authors show that doubly mutants for Adamts10 and Adamts6 suffer more severer skeletal developmental defects than single mutants. The authors also show that Adamts6 directly binds to Fbn2 and cleaves Fbn2 via a biochemical approach. Lastly, analysis of compound mutants of Adamts6 and Fbn2 or Fbn1 demonstrates that prevention of Fbn2 accumulation, but not Fbn1, rescues that Adamts6 KO skeletal phenotype and the reversed the aberrant BMP signaling. 

This study elegantly shows the physiologic importance of Fbn2 proteolysis by Adamts6 in skeletal development. On the other hand, the role of Adamts10 in Fbn2 proteolysis was previously demonstrated in vitro, and the mouse phenotype of a mutant Adamts10, including accumulation of Fbn2 and dampened BMP signaling with normal TGF signaling, was previously reported. In this regard, the major findings in this paper are somewhat expected. 

The strength of this paper is that it demonstrates the critical role of Adamts6 in Fbn2 proteolysis in skeletal development by combining mouse genetic, cell biological, and biochemical approaches. The experiments were conducted rigorously, and the conclusions are solid. The major weakness is that because the results are somewhat expected based on the knowledge from previous studies, this work gives an impression of being incremental. Also, as it is narrowly focusing on the role of Adamts6 on Fbn2 proteolysis, the significance of the findings does not seem very clear to a broader audience. 

The experiments were performed very well and the presented evidence supports the authors' conclusions.

Reviewer #3 (Public Review): 

This paper explores the shared and unique functions of the structurally related proteases ADAMTS6 and ADAMTS10 in the developing cartilage growth plate, using genetic and biochemical approaches to show they are both required, but use distinct mechanisms to promote the switch from embryonic prevalence of fibrillin-2 microfibrils to postnatal prevalence of fibrillin-1. This conclusion is supported by a phenotypic analysis documenting an essential function for ADAMTS6 in cartilage and a significant genetic interaction between Adamts6 and Adamts10 in this tissue that is clear from the greater severity of defects in Adamts6/10 double mutants. The phenotypic analysis of the growth plate defects is not comprehensive, but clearly documents elevated fibrillin-2 levels in the cartilage matrix, consistent with an essential role for both proteases in clearance of fibrillin-2. These findings are coupled with mechanistic in vitro and biochemical studies showing that ADAMTS6 directly interacts with and cleaves Fibrillin-1 and -2. Most significantly, the work demonstrates that Fibrillin-2 is an essential substrate of ADAMTS6 because the skeletal defects in Adamts6-/- mice are significantly rescued in Adamnts6-/-;Fbn2+/- mutants. Thus, the Adamts6 mutant phenotype can be largely attributed to inappropriate accumulation of Fibrillin-2. The paper also investigates whether the greater severity of skeletal manifestations in Adamts6-/- mice compared to Adamts10-/- mice may be due in part to compensatory transcriptional upregulation of Adamts6 in Adamts10-/- mice. However, although the authors argue that transcriptional upregulation of Adamts6 contributes to the milder skeletal phenotype of Adamts10 mutants, whether the observed upregulation in mRNA levels translates to elevated ADAMTS6 protein levels is unknown, and whether the less severe Adamts10 phenotype might reflect the presence of different ADAMTS protein that has similar function to ADAMTS10, is unclear. Nonetheless, the data represent an important contribution to our understanding of the regulation of fibrillin microfibril deposition and clearance.

Reviewer #1 (Public Review):

The manuscript addresses the question whether low P53 activation has a functional role in controlling normal cell function, with a focus here on correctly executing the cell cycle. To do so, the authors study cells either with stress (exposure to MNNG) or without, and compare the response to both cases for wild-type cells (which presumably have low P53 activation) and p53 siRNA cells (which presumably lack P53 activation altogether). On the technological side, the authors use automated cell tracking to measure individual cell lineages, manual annotation of specific cell division defects, and use complex simulations that generate virtual data from their measured single-cell data, to study the impact of their observations on scenarios that were not or could not be studied experimentally.

Major strengths:

- The experimental dataset gives a very detailed view on cell cycle dynamics and its dependence on P53, as it includes both cell lineage data and data on the occurrence of relevant, but infrequent events, such as cell death or cell fusion, that impact population dynamics.<br /> - This data set is used to reveal patterns in such defects that would not be visible without such single-cell analysis. For instance, it reveals a link between cell fusion, which occurs preferentially between siblings, and subsequent multipolar division.

Major weaknesses:

- The measure lineage data is subsequently used to generate simulation data, including for scenarios that were not measured experimentally. A major weakness was that the simulation algorithm was both highly complex, but insufficiently explained. As a consequence, it was not clear what the underlying assumptions of the simulations were and how these assumptions were based on and/or constrained by the experiments.<br /> - The single-cell analysis, including measuring lineages, by itself is not cutting-edge and has been done before, and so the novelty should be in the analysis. However, in many cases, the resulting data is presented in a manner that does not rely on the single-cell tracking (e.g. total cell number vs time in Fig. 2, average frequency of events in Fig. 4).<br /> - The impact of p53 was only assessed on level of differences between experimental conditions (p53 siRNA or not), but p53 levels themselves were not measured and therefore not incorporated in the single-cell analysis. In general, differences between wild-type and p53 siRNA data were small, while cell-to-cell variability in p53 knock-down appears high (as judged by Supplementary Fig. 4). This leaves open whether the relatively minor difference between wild-type and p53 siRNA cells reflects variability in p53 knockdown between cells, which is currently not directly assessed.

Reviewer #2 (Public Review):

In the manuscript, "Empirical single-cell tracking and cell-fate simulation reveal dual roles of p53 in tumor suppression," Rancourt et al examine how the tumor suppressor gene p53 can protect a cell in normal conditions and how its failure can lead to tumor proliferation. They use single-cell tracking of thousands of cancer cells imaged over several days in vitro with and without carcinogenic drug application, as well as extrapolated simulations of different scenarios and outcomes.

Overall, the study is interesting and its strengths lie in the identification of the cellular consequences of p53 function and malfunction in cell proliferation, in the presence and absence of a carcinogenic insult. The paper makes good use of single-cell analysis and simulations. The study's main weakness is the lack of empirical evidence from the simulation predictions of biology, and that the cellular consequences of p53 function were predictable and mostly confirmatory. The study uses novel methods and excellent use of single-cell tracking and imaging of cell fate, and the simulations quantify specific conditions and make interesting and testable predictions of biological function.

Reviewer #3 (Public Review):

The authors use a combination of cell lineage tracking and knockdown approaches to compare cell proliferation in cells expressing wild-type p53 (wt p53) compared to cells where p53 is knocked down using siRNA (siRNA p53). The imaging is based on differential interference contrast (DIC), a white light technique that avoids the use of fluorescent probes. Cellular levels of the p53 protein were lowered using siRNA, and notably, the the internalization of the siRNA particles in the cell was directly observed. These results were correlated with p53 expression in cells biochemically. Control and knockdown cells were then studied under basal and genotoxic conditions. Through a detailed analysis of cell fate and simulations, the authors conclude that cells expressing less p53 under baseline conditions can initiate tumour expression when tissue that contains a mixture of cells containing both normal and low levels of p53 are exposed to genotoxic stress.

Through direct visualization and tracking of individual cell fate, including cell division, cell growth and cell death, the authors observe track the lineages and events during proliferation. Through simulations, they can analyze the data in a way that decouples the effects of cell division on proliferation from cell death, and they use this approach in different combinations as they test and simulate other conditions. For example, although the siRNA cells yield more cells per division on average, their proliferation is about the same as that observed for wt p53 cells. Through simulations, the authors demonstrate that this arises because that siRNA p53 cells undergo cell death more frequently. When exposed to genotoxic stress, the situation is reversed, with siRNA p53 cells proliferating faster due to increased cell death in the wt p53 cells. The authors conclude that the results demonstrate that homeostatic levels of p53 are relevant in cancer progression, with lower levels being a risk. Yet it is unclear how these results can be generalized because the authors only studied one cell line. The results are not compared to other cell lines or primary cells, in terms of baseline expression of p53. In addition, it is unclear how this model is superior to testing homeostatic p53 compares to models that use mutated p53.

Strengths

Generally, it is easier to track cell lineages automatically from using fluorescence micrographs because these data are easier to segment from the background. Yet fluorescence probes have the potential to create off-target effects that may affect biological functions. For this reason, the authors' methodology would be of interest to anyone studying cell proliferation at the single cell level. The tools described, including the DIC tracking software and the simulation algorithms would be useful additions to the biologist's toolkit. The direct visualization of siRNA transfection agents through DIC, and its integration with western blotting is novel, and the authors may consider preparing a protocol or methods paper that describes this in more detail, as it may be useful for trouble-shooting when encountering difficulties with siRNA transfections.

The level of dedication and attention to detail in reporting the full cell lineage results and the details of the simulation algorithm is outstanding. Together, the manual and corresponding automatically tracked data are invaluable ground truth data sets for any researcher interested in modelling and simulating cell proliferation near the point of confluency. The use of white-light imaging is refreshing, as many of us in the field default to fluorescence imaging, which has the potential to interfere with cell proliferation. Overall, the approach is innovative by extracting the most information possible from optical imaging data sets, in the less invasive way possible.

Reviewer #1 (Public Review):

Technically, the paper from the Kaplan rests on solid ground. An array of mutations and transgenic lines are used in the study the Shank gene, and are nicely documented. The electrophysiological assessment of ionic currents in C. elegans muscle is clear and well documented. Finally, the light-level protein localization analyses are clear and well documented. The authors cleanly define a set of phenotypes caused by mutations in the Shank gene, influencing muscle action potentials in C elegans.

Reviewer #2 (Public Review):

The authors use C. elegans to explore the relationship between shank, CaV1 (Ca) channels, and BK (slo) calcium-dependent K channels in controlling muscle excitability. They use a range of genetic approaches to mutate or knock out one or more of these players and assess the impact on muscle action potential generation and slo currents. Their data show that shank controls AP width and plateau potential generation (pp) through Slo channels, and this effect is cell-autonomous in muscle. They go on to suggest this effect is mediated through CaV1-slo coupling, by using previously characterized mutant that reduce this binding. Because these mutations may affect binding with other partners, these experiments do not unequivocally implicate direct binding between these three players; however, use of fast and slow Ca buffers also suggests that shank keeps CaV1 and slo in close association, presumably allowing Ca influx through CaV1 channels to effectively activate slo channels. Finally, they show that overexpression of shank has a similar impact on excitability as reduction; this is somewhat puzzling and not further explored mechanistically, but is interesting given gene dosage effects of shanks in humans.

Reviewer #3 (Public Review):

Gao et al. present a nice set of data, using electrophysiology and molecular genetics, to address the function of C. elegans Shank (shn-1) in shaping muscle action potentials. Using genome-edited Cre-dependent deletion and expression of SHN-1, they show that removal of shn-1 specifically in body muscle widens the duration of action potentials and increases prolonged depolarization events known as plateau potentials (PP). They provide new evidence that SHN-1 couples the activity of the calcium channel EGL-19 to that of the BK potassium channels SLO-1/2. They additionally reveal that action potentials are sensitive to SHN-1 dosage. The experiments are generally conducted rigorously, and conclusions are stated appropriately. The findings offer insights into how human Shank misexpression might contribute to neurological disorder.

Reviewer #1 (Public Review):

Nguyen Ba and coworkers report the development of a clever novel approach for QTL mapping in budding yeast, dubbed "BB-QTL". In brief, they use batches of barcoded yeasts to generate very large barcoded F1 libraries (100,000 cells), followed by a Bar-Seq approach to map the fitness of these individuals and a clever low-coverage whole-genome sequencing coupled to background knowledge of the parental sequences to map their respective genotypes. A custom analysis pipeline then allowed predicting QTLs as well as possible epistatic interactions for a set of 18 phenotypes.

The novel technology expands the precision and power of more traditional approaches. The results mainly confirm previous findings. S. cerevisiae phenotypes are typically influenced by many different QTLs of different nature, including coding and noncoding variation; with coding and rare variants often having a larger effect. Moreover, several QTLs located in a set of specific genes like MKT1 and IRA2, were confirmed to influence multiple phenotypes (pleiotropy). Apart from confirming previous findings, the increased power of BB-QTL does offer the advantage of having lower error rates and higher power to detect specific mutations as drivers of a QTL, including some with only small effect sizes. Together, this yields a more complete and precise view of the QTL landscape and, most importantly, confirms widespread epistatic interactions between the different QTLs. Moreover, now that the barcoded pools have been developed, it becomes relatively easy to test these in other conditions. On the other hand, the power to detect many novel (industrially-relevant) QTLs is likely limited by the inclusion of only two parental strains, one being the lab strain BY4741.

Reviewer #2 (Public Review):

Ngyuyen Ba et al. investigated the genetic architecture of complex traits in yeast using a novel bulk QTL mapping approach. Their approach takes advantage of genetic tools to increase the scale of genetic mapping studies in yeast by an order of magnitude over previous studies. Briefly, their approach works by integrating unique sequenceable barcodes into the progeny of a yeast cross. These progeny were then whole genome sequenced, and bulk liquid phenotyping was carried out using the barcodes as an amplicon-based read-out of relative fitness. The authors used their approach to study the genetic architecture of several traits in ~100,000 progeny from the well-studied cross between the strains RM and BY, revealing in greater detail the polygenic, pleiotropic, and epistatic architecture of complex traits in yeast. The authors developed a new cross-validated stepwise forward search methodology to identify QTL and used simulations to show that if a trait is sufficiently polygenic, a study at the scale they perform is not sufficiently powered to accurately identify all the QTL. In the final section of the paper, the authors engineered 6 individual SNPs and 9 pairs of RM SNPs on the BY background, and measured their effects in 11 of the 18 conditions used for QTL discovery. These results highlighted the difficulty of precisely identifying the causal variants using this study design.

The conclusions in this paper are well supported by the data and analyses presented, but some aspects of the statistical mapping procedure and validation experiments deserve further attention.

In their supplementary section A.3-1.5 the authors perform QTL simulations to assess the performance of their analysis methods. Of particular interest is the performance of their cross-validated stepwise forward search methodology, which was used to identify all the QTL. However, a major limitation of their simulations was their choice of genetic architectures. In their simulations, all variants have a mean effect of 1% and a random sign. They also simulated 15, 50, or 150 QTL, which spans a range of sparse architectures, but not highly polygenic ones. It was unclear how the results would change as a function of different trait heritability. The simulations should explore a wider range of genetic architectures, with effect sizes sampled from normal or exponential distributions, as is more commonly done in the field.

In this simulation section, the authors show that the lasso model overestimates the number of causal variants by a factor of 2-10, and that the model underestimates the number of QTL except in the case of a very sparse genetic architecture of 15 QTL and heritability > 0.8. This indicates that the experimental study is underpowered if there are >50 causal variants, and that the detected QTL do not necessarily correspond to real underlying genetic effects, as revealed by the model similarity scores shown in A3-4. This limitation should be factored into the discussion of the ability of the study to break up "composite" QTL, and more generally, detect QTL of small effect.

In section A3-2.3, the authors develop a model similarity score presented in A3-4 for the simulations. The measure is similar to R^2 in that it ranges from 0 to 1, but beyond that it is not clear how to interpret what constitutes a "good" score. The authors should provide some guidance on interpreting this novel metric. It might also be helpful to see the causal and lead QTLs SNPs compared directly on chromosome plots.

The authors performed validation experiments for 6 individual SNPs and 9 pairs of RM SNPs engineered onto the BY background. It was promising that the experiments showed a positive correlation between the predicted and measured fitness effects; however, the authors did not perform power calculations, which makes it hard to evaluate the success of each individual experiment. The main text also does not make clear why these SNPS were chosen over others-was this done according to their effect sizes, or was other prior information incorporated in the choice to validate these particular variants? The authors chose to focus mostly on epistatic interactions in the validation experiments, but given their limited power to detect such interactions, it would probably be more informative to perform validation for a larger number of individual SNPs in order to test the ability of the study to detect causal variants across a range of effect sizes. The authors should perform some power calculations for their validation experiments and describe in detail the process they employed to select these particular SNPs for validation.

In section A3-1.4, the authors describe their fine-mapping methodology, but as presented is difficult to understand. Was the fine-mapping performed using a model that includes all the other QTL effects, or was the range of the credible set only constrained to fall between the lead SNPs of the nearest QTL or the ends of the chromosome, whichever is closest to the QTL under investigation? The methodology presented on its face looks similar to the approximate Bayes credible interval described in Manichaikul et al. (PMID: 16783000). The authors should cite the relevant literature, and expand this section so that it is easier to understand exactly what was done.

The text explicitly describes an issue with the HMM employed for genotyping: "we find that the genotyping is accurate, with detectable error only very near recombination breakpoints". The genotypes near recombination breakpoints are precisely what is used to localize and fine-map QTL, and it is therefore important to discuss in the text whether the authors think this source of error impacts their results.

The use of a count-based HMM to infer genotypes has been previously described in the literature (PMID: 29487138), and this should be included in the references.

Reviewer #1 (Public Review):

Here Laundon et. al report a cellular "atlas" of the model chytrid Rhizoclosmatium globosum. The data presented include beautiful and informative 3D reconstructions of all four key life stages of this species, as well as transcriptional profiling of matched samples and analysis of lipid compositions. The data were collected from multiple biological replicates and represent a clearly important resource for the community. With this work, the goal of the authors is to link structural descriptions of chytrid morphology with molecular understanding: this is something that the field needs. The authors describe results in three areas that are very interesting to the field. Unfortunately, the evidence provided for these findings does not always support their conclusions. Additionally, discussion of the literature is insufficient as previous work provides crucial context for interpreting the data presented.

Major points

1. Zoospores in several chytrids have been shown to be transcriptionally and translationally inactive, this means that the distribution of transcripts are maternally allocated. Although the authors do cite two papers on the topic in the discussion, this is a fundamental concept that might not be in the mind of non-specialist readers and the authors need to introduce and discuss from the beginning (see PMID: 4412066, 1259436, 3571161), as it provides key context for their finding that germlings have a wider range of transcriptional activity as this is consistent with Rg also being transcriptionally silent in the zoospore state. Finally, the language used to describe transcripts found in zoospores (the manuscript refers to them "expressing" particular genes) is confusing given this context.

2. The authors correlated structural changes observed with general KEGG pathway profiles obtained from transcriptomics. Unfortunately it is hard to pin down exactly what the authors are trying to say about this data because their observations are not placed with precision in the context of what is already known about chytrid development, and KEGG pathways are too broad to be very informative. Drawing inferences about chytrid biology from broad KEGG categories and link them to structural observation is not possible with more detailed molecular analysis. This comes up multiple times: (i) Correlation between an increase in endomembrane structures in a compartment and enrichment of KEGG categories of protein processing and ER etc is not enough to link these endomembrane systems with ER. This requires more direct evidence. (ii) High dynamic activity and endomembrane density in the apophysis is not evidence enough by itself to support the claim of the "apophysis acting as a cellular junction that regulates intracellular traffic." (iii) Although different lipid composition between zoospore and germling, and differences in KEGG categories of peroxisome activity on the other suggest important lipid metabolic changes, these correlations are is not hard enough evidence for the authors to call this process as a "biological characteristic" of the transition from zoospore to germling.

3. The claim that zoospores inside the sporangium undergo phagocytosis is not sufficiently supported by the data presented. To date there is only one case in which it a fungus undergoes a process akin to phagocytosis (i.e. Rozella), and finding a phagocytic fungus would be a very exciting result. Unfortunately, the authors provide no direct evidence to support this specific claim as (i) there are many ways one could imagine to explain the shapes seen in the EM data (perhaps the zoospores are squeezed around the the objects), and classic work on Allomyces and Blastocladiella zoosporogenesis indicates that cleavage vesicles can be orderly or very irregular before they align in continuous plates (sometimes concomitant with formation of ribosome aggregates), and that these cleavage planes are nearly complete, but not complete yet. (ii) The genes discussed are not specific to phagocytosis, but are used for a wide variety of other functions. Moreover, the authors appear to equate endocytosis and phagocytosis, and although there is some overlap in the proteins used for these processes, they are not equivalent.

4. Although the author's findings about the complex endomembrane system in Rg apophysis is interesting, the details of the images provided do not support their interpretation of it being a "distinct subcellular structure". Such claims require detailed imaging of the "pseudo-septum", similar to what has been shown for "plasmodesmata" in Entophlyctis and Blastocladiella.

Reviewer #2 (Public Review):

The authors are developing and investigating a molecular atlas of the cell for the chytrid Rhizoclosmatium globosum. They are linking transcriptome and lipidome information data to cellular biology that can be observed through SBF-SEM. The detailed investigation allowed a development of a cell atlas and interpretation of cell wall and organelle dynamics.

The authors successfully explored several hypotheses about development in chytrids including that zoospores are provisioned with maternal mRNA. They interpret the composition of spores to include both essential machinery and instructions for cellular replication but also have more host- or substrate-interaction products primed for the cell to condition development on external signals.

They also explore whether cell wall genes are expressed in a fashion that links to the cell wall-less spore stage, and seem to indicate there are at least one chitin synthase with high levels indicating preparation for dynamic growth and wall formation.

The RNASeq analysis/GO enrichment pointed to secondary metabolism enrichment in some of the comparisons, but there is little discussion of what types of genes these might be in the manuscript. Are these NRPS and siderophore or other product producing genes that contribute to that enrichment category?

The work will have an important impact in the field of cell biology of chytrids but also broadly to Fungi and perhaps also comparative biology of Opisthokonts. The detailed reconstruction and examination of the cellular structures in this lineage should be informative to how transitions occurred in the development of septa in other flagellated lineages (eg Blastocladiomycota) and in Dikarya. The development from encysting zoospore to thallus is clearly complex and this study gives a high resolution and a dynamic examination of the processes.

Reviewer #1 (Public Review):

The authors show an important role of an RNA-binding protein (RBP), YTHDF2 in the accumulation of plasma cells. In addition, by a CRISPR/Cas9 knockout screening of RBPs, the authors suggest that some RBPs are involved in plasma cell differentiation. The roles of RBPs in a lymphocyte differentiation system are very interesting. The methods to detect germinal centre B cells and plasma cells could be improved.

Reviewer #2 (Public Review):

Turner et al. investigate the role for RNA binding proteins (RBPs) in regulating B cell to plasma cell differentiation in mice. They find sets of RBPs that control distinct phases of B cell differentiation including proliferation, survival, and the terminal differentiation of CD138+ plasma cells. They find only a few RBPs promote proliferation and hundreds of RBPs that control terminal differentiation. Follow up studies confirm the effect for select RBPs and the authors focus on the YTHDF2 gene which recognizes N6-methyladenosine in RNA. Using genetic deletion and bone marrow chimera models, the authors demonstrate a role for YTHDF2 in regulating plasma cell formation in response to NP-KLH immunization in both the spleen and bone marrow. Competitive bone marrow chimeras show that germinal centers and early B cell activation are normal in the absence of YTHDF2, but a significant decrease in bone marrow plasma cells is observed. The authors then using m6A-eCLIP and performed RNA-seq on the same cell types to define m6A modified transcripts. Contrary to the hypothesis, no enrichment for m6A-modified transcripts was observed for genes that repressed plasma cell formation and were predicted to be YTHDF2 targets.

In its current form, the conclusions of the paper are not fully supported by the data. The number of samples per experimental group and whether experiments were reproducible across independent groups is not clear and needs to be clarified.

Strengths:<br /> The area of RBP biology is underexplored in immune system function and the authors establish a powerful CRISPR/Cas9 sgRNA pool that will be a resource in the B cell field. Additionally, the use of sophisticated tools such as the two bone marrow chimera models, the tracking of NP-specific immune responses following NP-KLH immunization, and mapping of m6A by eCLIP allows for clear conclusions to be made.

Weaknesses:<br /> It is not clear if sufficient replicates or statistics were used to demonstrate reproducibility and support the conclusions. For example, experiments in Fig 1C and 1F are critical to independently validate the results of the CRISPR/Cas9 screen, yet only 2-3 data points are presented, and no indication is given if the experiments were independently replicated across more than one cohort. Also, the same concern of independent replicates is raised for the data in Fig 2 and 3. Additionally, no evidence is provided that the ratios of Cas9+/Cas9- cells are statistically different from the NT controls. The fold-changes are small compared to the NT sample, and without flow cytometry data showing the percentage of CD138+ cells it is difficult to interpret what the true effect size is. Without this information, the authors conclusion that the CCR4-CNOT complex plays any role in plasma cell differentiation is not well supported.

The data do not support the authors conclusion that IRF4 only affects B cell differentiation. IRF4 falls on the diagonal in the scatter plot in Fig 1D, indicating it also affects proliferation/survival. In fact, IRF4 has been previously shown to regulate B cell proliferation (Sciammas et al. 2006 Immunity) and differentiation to plasma cells.

The validation of YTHDF2 and its role in plasma cell differentiation but not prior differentiation stages is a valuable section of the study. However, there are concerns about using only flow cytometry to measure very rare populations of plasma cells. From the data presented, roughly 8-10 plasma cells were counted per million cells.

Reviewer #3 (Public Review):

The mammalian genome contains thousands of RNA binding proteins. However, the importance of these proteins in regulating plasma cell differentiation is largely unknown. The authors sought to identify RNA binding proteins regulating the differentiation of plasma cells. They achieved this aim by using a Crispr-Cas9 screen to identify 292 RNA binding proteins that regulate the differentiation of CD138+ cells in vitro. This study effectively demonstrated the utility of Crispr-Cas9 screens in identifying factors regulating B cell differentiation.

One limitation of this study is that the RNA binding proteins identified as regulating the differentiation of CD138+ cells in vitro may not necessarily have the same role in vivo. While the authors validated that the RNA m6A binding protein YTHFD2 regulated plasma cell differentiation following protein immunization, additional work will be required to determine the relevance of other RNA binding proteins identified in their screen. An additional limitation of this study is that the authors did not determine the mechanisms by which YTHFD2 promotes plasma cell differentiation. This lack of mechanistic insight limits the utility of this study in providing a conceptual advance in the understanding of the processes governing plasma cell differentiation. However, the results of their screen will still likely be a useful resource for the future work seeking to more precisely understand how RNA binding proteins regulate B cell differentiation.

Reviewer #1 (Public Review):

Overview

This is a well-conducted study and speaks to an interesting finding in an important topic, whether ethological validity causes co-variation in gamma above and beyond the already present ethological differences present in systemic stimulus sensitivity.

I like the fact that while this finding (seeing red = ethnologically valid = more gamma) seems to favor views the PI has argued for, the paper comes to a much simpler and more mechanistic conclusion. In short, it's good science.

I think they missed a key logical point of analysis, in failing to dive into ERF <----> gamma relationships. In contrast to the modeled assumption that they have succeeded in color matching to create matched LGN output, the ERF and its distinct features are metrics of afferent drive in their own data. And, their data seem to suggest these two variables are not tightly correlated, so at very least it is a topic that needs treatment and clarity as discussed below.

Minor concerns

In generally, very well motived and described, a few terms need more precision (speedily and staircased are too inaccurate given their precise psychophysical goals)

I got confused some about the across-group gamma analysis:

"The induced change spectra were fit per participant and stimulus with the sum of a linear slope and up to two Gaussians." What is the linear slope?

To me, a few other analyses approaches would have been intuitive. First, before averaging peak-aligned data, might consider transforming into log, and might consider making average data with measures that don't confound peak height and frequency spread (e.g., using the FWHM/peak power as your shape for each, then averaging).

Moderate

I. I would like to see a more precise treatment of ERF and gamma power. The initial slope of the ERF should, by typical convention, correlate strongly with input strength, and the peak should similarly be a predictor of such drive, albeit a weaker one. Figure 4C looks good, but I'm totally confused about what this is showing. If drive = gamma in color space, then these ERF features and gamma power should (by Occham's sledgehammer...) be correlated. I invoke the sledgehammer not the razor because I could easily be wrong, but if you could unpack this relationship convincingly, this would be a far stronger foundation for the 'equalized for drive, gamma doesn't change across colors' argument...(see also IIB below)...

...and, in my own squinting, there is a difference (~25%) in the evoked dipole amplitudes for the vertically aligned opponent pairs of red- and green (along the L-M axis Fig 2C) on which much hinges in this paper, but no difference in gamma power for these pairs. How is that possible? This logic doesn't support the main prediction that drive matched differences = matched gamma...Again, I'm happy to be wrong, but I would to see this analyzed and explained intuitively.

II. As indicated above, the paper rests on accurate modeling of human LGN recruitment, based in fact on human cone recruitment. However, the exact details of how such matching was obtained were rapidly discussed-this technical detail is much more than just a detail in a study on color matching: I am not against the logic nor do I know of a flaw, but it's the hinge of the paper and is dealt with glancingly.

A. Some discussion of model limitations

B. Why it's valid to assume LGN matching has been achieved using data from the periphery: To buy knowledge, nobody has ever recorded single units in human LGN with these color stimuli...in contrast, the ERF is 'in their hands' and could be directly related (or not) to gamma and to the color matching predictions of their model.

Reviewer #2 (Public Review):

The major strengths of this study are the use of MEG measurements to obtain spatially resolved estimates of gamma rhythms from a large(ish) sample of human participants, during presentation of stimuli that are generally well matched for cone contrast. Responses were obtained using a 10deg diameter uniform field presented in and around the centre of gaze. The authors find that stimuli with equivalent cone contrast in L-M axis generated equivalent gamma - ie. that 'red' (+L-M) stimuli do not generate stronger responses than 'green (-L+M). The MEG measurements are carefully made and participants performed a decrement-detection task away from the centre of gaze (but within the stimulus), allowing measurements of perceptual performance and in addition controlling attention.

There are a number of additional observations that make clear that the color and contrast of stimuli are important in understanding gamma. Psychophysical performance was worst for stimuli modulated along the +S-(L+M) direction, and these directions also evoked weakest evoked potentials and induced gamma. There also appear to be additional physiological asymmetries along non-cardinal color directions (e.g. Fig 2C, Fig 3E). The asymmetries between non-cardinal stimuli may parallel those seen in other physiological and perceptual studies and could be drawn out (e.g. Danilova and Mollon, Journal of Vision 2010; Goddard et al., Journal of Vision 2010; Lafer-Sousa et al., JOSA 2012). Similarly, the asymmetry between +S and -S modulation is striking and need better explanation within the model (that thalamic input strength predicts gamma strength) given that +S inputs to cortex appear to be, if anything, stronger than -S inputs (e.g. DeValois et al. PNAS 2000).

My only real concern is that the authors use a precomputed DKL color space for all observers. The problem with this approach is that the isoluminant plane of DKL color space is predicated on a particular balance of L- and M-cones to Vlambda, and individuals can show substantial variability of the angle of the isoluminant plane in DKL space (e.g. He, Cruz and Eskew, Journal of Vision 2020). There is a non-negligible chance that all the responses to colored stimuli may therefore be predicted by projection of the stimuli onto each individual's idiosyncratic Vlambda (that is, the residual luminance contrast in the stimulus). While this would be exhaustive to assess in the MEG measurements, it may be possible to assess perceptually as in the He paper above or by similar methods. Regardless, the authors should consider the implications - this is important because, for example, it may suggest that important of signals from magnocellular pathway, which are thought to be important for Vlambda.

Reviewer #3 (Public Review):

This is an interesting article studying human color perception using MEG. The specific aim was to study differences in color perception related to different S-, M-, and L-cone excitation levels and especially whether red color is perceived differentially to other colors. To my knowledge, this is the first study of its kind and as such very interesting. The methods are excellent and manuscript is well written as expected this manuscript coming from this lab. However, illustrations of the results is not optimal and could be enhanced.

Major

The results presented in the manuscript are very interesting, but not presented comprehensively to evaluate the validity of the results. The main results of the manuscript are that the gamma-band responses to stimuli with absolute L-M contrast i.e. green and red stimuli do not differ, but they differ for stimuli on the S-(L+M) (blue vs red-green) axis and gamma-band responses for blue stimuli are smaller. These data are presented in figure 3, but in it's current form, these results are not well conveyed by the figure. The main results are illustrated in figures 3BC, which show the average waveforms for grating and for different color stimuli. While there are confidence limits for the gamma-band responses for the grating stimuli, there are no confidence limits for the responses to different color stimuli. Therefore, the main results of the similarities / differences between the responses to different colors can't be evaluated based on the figure and hence confidence limits should be added to these data. It is also not clear from the figure legend, from which time-window data is averaged for the waveforms.

The time-resolved profile of gamma-power changes are illustrated in Fig. 3D. This figure would a perfect place to illustrate the main results. However, of all color stimuli, these TFRs are shown only for the green stimuli, not for the red-green differences nor for blue stimuli for which responses were smaller. Why these TFRs are not showed for all color stimuli and for their differences?

Reviewer #1 (Public Review):

Psychiatric symptoms in Parkinson's disease are debilitating, but there are few treatments that effectively reduce these symptoms long-term. The mechanisms that cause psychiatric symptoms in Parkinson's disease are unknown. However, it has been known for decades that abnormal alpha-synuclein is found in the amygdala, a brain region important for the control of emotions. Nagaraj et. al. present an article in which they attempt to characterize the differences in α-synuclein colocalization with vGluT1+ compared to vGluT2+ terminals in the BLA of a PFF mouse model. They successfully demonstrate convincing data that points to the preferential association of α-synuclein with vGluT1+ puncta and not vGluT2+ puncta. The authors also demonstrate that PFFs promote short-term depression of cortico-BLA synapses in response to repetitive stimuli which does not occur in vGluT2+ terminals.

Clearly differentiating the association of α-synuclein with different glutamatergic terminals and cortical or thalamic projections, and the subsequent effect of abnormal α-synuclein and how it affects transmission in the BLA is novel and points to mechanisms of differential vulnerability to inclusions in different neuronal bodies.

This study is one of the first to use electrophysiology to show that abnormal alpha-synuclein contributes to defects in the amygdala in Parkinson's disease. The study also pinpoints cortical-amygdala projections as the culprit in amygdala dysfunction. Therefore this study has a major impact in the field by determining how abnormal amygdala function caused by pathologic alpha-synuclein can potentially cause psychiatric symptoms in Parkinson's disease.

The main weakness of the study is the lack of mechanism. Although the authors attempt to show that loss of synuclein in mice injected with PFFs is responsible for the amygdala defects, the data are insufficient to make this conclusion.

Reviewer #2 (Public Review):

The data presented are clear and of high quality. The conclusion that alpha-synuclein aggregation and corresponding synaptic dysfunction preferentially occurs in vGluT1 expressing cortical inputs (as opposed to vGluT2 expressing thalamic inputs) to the BLA is convincing, but a few additional clarifications and experiments would greatly help describe the mechanism of synapse dysfunction. Overall this manuscript provides helpful insight into the circuit dysfunctions that may contribute to non-motor psychiatric symptoms that commonly occur in Parkinson's disease.

1. The BLA is a relatively large structure, and the labeled terminal fields of cortical and thalamic inputs (figure 2) don't show matching patterns. It would be helpful to clarify where in the BLA recordings were made (and where high mag images in figure 1 were taken from).<br /> 2. The short-term plasticity experiments shown in figure 4 are informative, but by themselves don't necessarily rule out post-synaptic mechanisms of adaptation. Since the mobilization of synaptic vesicles is likely involved, it would be helpful to also look at the effect of PFFs on release probability using pared pulse ratios.<br /> 3. PPFs reduce cortico-BLA EPSC amplitudes but not thalamo-EPSC amplitudes in response to single electrical and optogenetic stimuli (figure 2). In figure 4, however, the starting amplitudes appear to be similar (at least in the exemplar traces). I'm assuming this is because stimulus intensities were adjusted to achieve a similar starting point? If so, are differences in short-term plasticity also observed if similar stimulus intensities are used?

Reviewer #3 (Public Review):

In this manuscript, the authors try to address whether glutamatergic axonal terminals are differentially impacted by a-syn aggregation, a key pathology seen in Parkinson's disease. Using a-syn PFF injection, and a-syn KO mice, the authors show a few interesting findings: 1. After a-syn PFF injection in the BLA, the strength of the cortical inputs was selectively reduced, while leaving thalamic inputs unaffected. 2. There is an interesting parallel finding on the release probability of cortical glutamatergic synaptic transmission after a-syn PFF injection and in the a-syn KO mice. The key findings are interesting, showing selective vulnerability of glutamatergic synapses, in which vGluT1+ terminals are more profoundly affected by a-syn PFF or loss of function.

However, mechanistically, the authors implied that a-syn PFF induced aggregation sequesters soluble a-syn, acting more similar to a-syn KO conditions. This does seem to be plausible for the enhancement of release probability. But, what would be responsible for the reduction of cortico-BLA synaptic transmission? Previous studies showed that there was no neurodegeneration one month after a-syn PFF injections in the cortex.

Do the authors imply that some vGluT1+ terminals are lost after a-syn PFF injection? The authors did not quantify the number of vGluT1+ puncta in the BLA after a-syn PFF injections.

The authors also used intrastriatal a-syn PFF injection as a comparison. However, the data were not shown in the manuscript. Because striatum also receives convergent cortical and thalamic inputs, it would strengthen the conclusion if the authors systematically investigate corticostriatal vs thalmostriatal terminals in parallel.

Reviewer #1 (Public Review): 

In this manuscript, Birak and colleagues report single cell resolution whole brain imaging experiments in immobilized C. elegans at different ageing stages. It was shown previously that, under similar imaging conditions, animals exhibit brain-wide activity correlations and vigorous neuronal dynamics that evolve on a low dimensional manifold, visualised by principal components analyses (PCA). The transitions in network activity states were previously interpreted as motor command states, that correspond to forward-, backward- and turning- motor commands. The authors show that with increased ageing, the frequency of these transitions declines, the manifolds in PCA states appear less organized and specifically negative pair-wise correlations are reduced. They interpret this result as an ageing associated excitatory/inhibitory (E/I) imbalance caused by a decline in inhibitory signalling. They provide some experiments (imaging in unc-2 mutants, and in the presence of a GABAR agonist), the results of which are consistent with this interpretation.

The work provides an impressive amount of whole brain imaging experiments, and to the best of my knowledge is the first single cell resolution whole brain imaging study in any organism that reports how whole brain dynamics change over the course of ageing. The findings are therefore of broader interest. However, I think that the conclusions from their analyses results are premature and allow for alternative interpretations, which can be addressed with additional more detailed analyses. The genetic and pharmacological manipulations do not fully phenocopy or compensate the age associated changes in neuronal dynamics, therefore these results rather provide only indirect evidence for their conclusions, which could be more critically discussed.

Reviewer #2 (Public Review):

The study in "Age-associated changes to neuronal dynamics involve a loss of inhibitory signaling in_C. elegans_" attempts a hard task: To quantify changes in brain dynamics during aging. They do so in the nematode _C. elegans_, which has the advantage of allowing `whole brain` imaging, i.e., imaging of a large number of neurons in the head ganglion of the worm. A major strength of the paper is its deep rooting in the literature which establishes the context for the study with respect to both aging studies in the nematode as well as from other species. 

Using statistical analyses of the temporal neuronal dynamics, the authors find that the brain dynamics slow down with age and the balance between excitation and inhibition changes. Overall, the study manages to effectively combine genetics and neural imaging to quantify changes in an aging brain. They convincingly demonstrate slowed dynamics as worms get older and suggest a role for the unc-2 /ced-4 pathway in changing the inhibitory connections. 

Yet, these data are difficult to interpret: Brain dynamics are measured by imaging an encoded indicator GCaMP. The expression level of this indicator is also subject to age-dependent changes in expression level and possibly changes in the intracellular environment (e.g., pH). As much of the quantification is sensitive to the signal-to-noise ratio of the measured signals, this confounds the results. Similarly, some measures are sensitive to the inherent autocorrelation of the signal, and the slow-down in dynamics over age is possibly sufficient to effect the observed changes in all other metrics they use to compare between neural activity under different conditions. 

The paper is interesting, in that it attempts to connect the dynamics and system organization of brain dynamics with a molecular mechanism that is active during aging. However, at this point the deeper interpretation of specific findings hinges on the metrics being independent of the measured signal, and this is not convincingly demonstrated.

Reviewer #3 (Public Review): 

Although whole brain activity imaging is recently available in C. elegans, it is still difficult to map each cell activity to known neuronal cell identity. Without cell ID information, we cannot interpret the activity data in the context of neuronal circuits or statistically discuss neuronal activities across multiple individuals. The present study provides an excellent representation of age-dependent changes in the whole brain activity by calculating three indices: angular change in PCA trajectory, power spectral densities, and neuron pair correlation, without cell identification. In addition, they found age-dependent changes in anticorrelated neuron pairs, while almost no change in positively correlated pairs was observed. Furthermore, by integrating genetic and pharmacological analyses, they showed that reduction of GABA signaling via the CaV2/caspase pathway may be the cause of the changes. Still, additional experiments are needed to show whether the observed whole brain activity changes are indeed controlled by these gene products.

Reviewer #3 (Public Review):

Mo et al., present novel and, in most parts conclusive data showing that the function of TRPV2 is tightly regulated by phosphorylation of specific tyrosine residues. 

In the hands of the authors, in vitro electrophysiological experiments revealed a remarkable sensitizing effect of high concentrations of extracellular Mg on ion-currents activated by 2-APB or high heat on rat TRPV2. This effect is reproducible in different cell types and did not seem to differ between recombinant channels and TRPV2 endogenously expressed in macrophages and neurons. While it remains a little unclear why, or with what hypothesis the authors conducted these initial experiments, they present a fair amount of evidence indicating that this Mg-induced effect is due to an indirect intracellular effect rather than a direct interaction with Mg on specific extracellular or intracellular sites of TRPV2. While Mg applied alone completely failed to activate TRPV2 examined in the whole-cell mode as well as in inside-out cell-free patches, it potentiated currents evoked by 2-APB or heat. In consecutive experiments, the authors present some evidence that Mg permeates through the pore of TRPV2 to increase phosphorylation of TRPV2. Using selective inhibitors of different kinases, mass spectrometry and shRNA-mediated knockdown, the authors present very convincing evidence that JAK1 is the main kinase responsible for this Mg-induced effect. Of note, inhibition of JAK1 has also reduced phagocytosis of BMDM cells. Given that TRPV2 is known to be important for phagocytosis, this finding indeed indicates that the degree of phosphorylation of TRPV2 is crucial for the function of macrophages - e.g. giving this in vitro study an exciting physiological relevance. Using site-directed mutagenesis, the tyrosine residues Y335, Y471 and Y525 were found to the key sites for JAK1-induced phosphorylation. Finally, the authors very elegantly demonstrate that the degree of TRPV2-phosphorylation also seems to be finely tuned by the phosphatase PTPN1. 

In most parts of this well-written study, the conclusions drawn by the authors are supported by conclusive data generated by state-of-art in vitro techniques. The findings are truly novel, and as very little is still known about the precise functional properties of TRPV2, the data presented here may have a substantial impact on future research on these ion channels. Saying this, some aspects might be worth clarifying or at least discussing. 

1. The underlying hypothesis leading the authors to challenge TRPV2 with high concentrations of extracellular Mg should be (better) stated. There are many ways to explore the role of protein phosphorylation in the activity of ion channels, the authors applied many elegant techniques in this study. The starting point with extracellular Mg is somewhat odd, giving the slight impression that the authors did not have protein phosphorylation in mind when they started this study. 

2. A key event of basically all functional data presented in this study is the likely permeation of Mg through the pore of TRPV2. However, ion permeability of TRPV2 is not very well explored in previous reports. Thus, permeation of Mg through TRPV2 should be convincingly demonstrated. 

3. It is easy to accept that probably many signalling pathways in different cell types expressing TRPV2 may regulate phosphorylation of TRPV2 via JAK1/PTPN1. Excessive concentrations of is artificial, well suitable for this mechanistic in vitro study. However, it may be possible to at least speculate which are the endogenously relevant pathways involving JAK1 and PTPN1.

Reviewer #1 (Public Review): 

The TRPV2 channels are expressed in multiple cell types and have been shown to be essential for the development and function of the heart and activation of phagocytosis by macrophages. TRPV2 channels are directly activated by heating, but this requires extreme temperatures outside the physiologically relevant range (> 50{degree sign}C). Although some mechanisms such as oxidative modification of methionine residues have been found to enable channel activation by heat at physiological temperatures, the endogenous mechanisms of TRPV2 channel activation remain largely unknown. Here Xiaoyi Mo and collaborators describe a novel mechanism of TRPV2 channel regulation by phosphorylation/dephosphorylation that could tune TRPV2 channel sensitivity to heat and other stimuli allowing its activation in rat bone marrow macrophages and potentially other cell types under physiological conditions. Using patch-clamp electrophysiology, the authors find that extracellular magnesium co-applied with the chemical agonist 2-APB slowly sensitizes responses to sub-saturating concentrations of 2-APB and heat in rat bone marrow-derived macrophages and DRG neurons that endogenously express TRPV2, as well as in heterologous expression systems. The authors find no effect when other divalent cations are applied from the extracellular side or when large concentrations of a divalent ion chelator are included in the intracellular solution, suggesting that the inward flow of magnesium in a whole cell is required to observe sensitization of the channels. The authors also find no sensitization when cytosolic ATP is substituted by a non-hydrolyzable analog, suggesting that a kinase might be responsible for the observed effect of magnesium on channel sensitization. By testing multiple pharmacological kinase inhibitors, the authors find that only a JAK1 kinase inhibitor ablates the sensitizing effect of magnesium. Using biochemical assays and mass-spectrometry, the authors provide further evidence that the presence of magnesium enhances TRPV2 channel phosphorylation, and show that siRNA-dependent knockdown of JAK1 kinase reduces magnesium-dependent sensitization in patch-clamp experiments and biochemical assays. Using site-directed mutagenesis, the authors identify three tyrosine residues responsible for the observed sensitizing effect. Finally, using pharmacological phosphatase inhibitors, siRNA-dependent knockdown of different phosphatases, and biochemical assays of phosphorylation, the authors find that protein tyrosine phosphatase non-receptor type 1 (PTPN1) is capable of de-phosphorylating TRPV2 channels to reverse sensitization caused by JAK1. The data presented are of high quality, and most conclusions are supported by the evidence provided. The findings are relevant, as they identify a novel regulatory mechanism that could enable TRPV2 channel activity under physiological conditions. 

One concern in the data is that application of 2-APB alone is much shorter than when co-applied with extracellular magnesium in all experiments in the manuscript (e.g. Figs. 1A, C, E, Fig. 1 Suppl. 1 A-C). Because the sensitization caused by 2-APB and magnesium takes > 100 s to develop, it is unclear from the data whether a longer stimulation with 2-APB alone would have some sensitizing effect. In general, the authors do not appear to account for the length of stimulation with 2-APB and magnesium in experiments where this is an important factor, such as the dose-response relations for magnesium or 2-APB (Fig. 1G and H), the comparison between different divalent cations (Fig. 1 Suppl. 2 A-H), the effects of kinase or phosphatase inhibitors (Figs. 2E and 5B), the effect of magnesium applied alone at different concentrations (Fig. 2A) or the effects of the various tyrosine mutants (Fig. 4A, G, H and I). 

Normalization of the current vs temperature relations and the quantitation of the threshold is not appropriate, because the apparent threshold depends on how data is normalized and different temperatures are used to normalize the data sets being compared (Figs. 1K, 5K, 6G, and Fig. 5 Suppl. 1G). 

The data provided to support that channels containing substitutions E609Q and E614Q have reduced permeability to divalent cations and magnesium specifically is not adequate. First, the calcium imaging data is largely qualitative and provides no information about magnesium permeability. It is also unclear why the authors utilized CBD instead of 2-APB. Direct measurement of the reversal potential with or without extracellular magnesium would be required. 

Regarding the in-vitro kinase assays (Fig. 3F), it is unclear why a smear is apparent on the third lane, and also limited information is provided regarding the mass-spectrometry results, making it hard for the general reader to assess these results. 

An important concern pertains to the phagocytosis assay. Because the authors are utilizing primary cells for this, it is unclear whether the effects of JAK1 inhibition on phagocytosis are specific to TRPV2 channel activity. These cells do express other types of ion channels, and the observed changes are not very robust. Importantly, the authors did not stimulate cells with magnesium and 2-APB, so it is unclear what other signal would be responsible for triggering TRPV2 channel activity and phagocytosis. 

Finally, the authors provide very limited discussion about how TRPV2 channel phosphorylation/de-phosphorylation could be modulated physiologically. To observe robust sensitization, the authors need to raise the extracellular magnesium concentration > 10 mM. It is unclear whether the resulting increases in cytosolic magnesium under these conditions could occur physiologically, or whether there are other signaling pathways regulating the function of JAK1 of PTPN1.

Reviewer #2 (Public Review): 

How TRPV2 is exactly integrated into signaling pathways of an organism is a relatively open question. There is no compelling evidence for a specific endogenous TRPV2 ligand, and its thermosensitive threshold at 52 degrees Celsius is nearly outside the practical range, where it could help an organism avoid noxious temperatures. These findings have prompted speculation that cell signaling pathways play an important part in the physiological role of the channel. The research article, "Tyrosine phosphorylation tunes chemical and thermal sensitivity of TRPV2 ion channel", by Mo et al. makes a significant contribution to our understanding of how TRPV2 function is integrated with cell signaling and its physiological role. By leveraging a newly discovered TRPV2 response to rapid recovery of cellular Mg2+ levels from low initial concentrations, the authors conduct a compelling study of how phosphorylation of TRPV2 at three amino acids determine the functional response of the channel in the physiology of bone-marrow-derived macrophages (BMDM). The research follows a logical progression of experiments to identify the key amino acids that are targeted for phosphorylation in this pathway and is a fine example of a clearly conceptualized and actualized study. In terms of the interpretation of the results, the importance of the phosphorylation events in TRPV2 in relation to physiological function is well supported by the evidence. However, some caution should be given to interpreting the Mg2+ dependent mechanism of TRPV2 modulation as having physiological relevance. The Mg2+ experiments with TRPV2 proceeded from an initial state with very low intracellular Mg2+ concentrations to relatively high levels (0.1-100 mM), and no experiments were conducted that explored Mg2+ concentration changes from nominal initial levels (1-5 mM).

Reviewer #2 (Public Review):

In this manuscript, the authors attempt to identify the essential role of Oct4 in mouse ESCs. They utilized existing Tet-OFF Oct4 ESCs and looked for changes in gene expression, enhancer RNA expression and enhancer chromatin accessibility in a timecourse while depleting Oct4. This led to the discovery that many enhancer elements bound by Oct4 showed downregulation of enhancer RNAs following loss of Oct4. This typically occurred before reduction of gene expression and before or in the absence of loss of chromatin accessibility. They further showed that Sox2 remained bound at high levels for some time at enhancers that proceeded to be inactivated, while enhancers at which Oct4 was not required for activity retained Sox2 binding throughout the timecourse.<br /> The in-depth analysis of nascent enhancer RNA expression and chomatin accessibility changes during a timecourse following loss of Oct4 are interesting and provide insights into the order of events that occur during enhancer decommissioning. However, there are several key limitations to this study with regard to identifying the primary function of Oct4 in the control of pluripotency, and as such several of the key conclusions may be somewhat overstated.

Reviewer #1 (Public Review):

While the work is of great interest, some of the key conclusions are not supported by the data as the manuscript stands, due to the progressive loss (rather than dissaperance) of Oct4 activity, the possible differences among regions that require or not Oct4 binding and the role of Sox2, which requires experimental support.

Reviewer #3 (Public Review):

In this work by Le Xiong et al., the authors focus on the role of Oct4 in activating transcription at target enhancers and genes and its ability to regulate chromatin accessibility. To do so, they used a previously established Tet-off Oct4 system to deplete Oct4 levels gradually over a period of 15 hours. They performed TT-seq and ATAC-seq experiments over this time frame with a time resolution of 3 hours. They found that eRNA transcription rapidly decreases in response to a decrease in Oct4 levels. Among the enhancers decreasing eRNA synthesis in response to a decrease in Oct4 levels, about half of them displayed a decrease in accessibility and the other half does not. They found that chromatin accessibility changes at loci that do decrease their accessibility in response to Oct4 knockdown are delayed as compared to changes in transcriptional activity. They also find that Sox2 occupancy is maintained or decreased at loci that do not change or decrease their accessibility in response to Oct4 knockdown, respectively. From these results, they conclude that Oct4 regulates transcriptional activity but is not critical for regulation of chromatin accessibility.

The major strengths of the paper is the high quality of the experiments that assess chromatin state and acute transcriptional changes using state of the art methods.<br /> The fine kinetics of transcriptional/chromatin accessibility changes upon Oct4 removal, and the detailed dissection of how different genomic loci are temporally affected by these changes is a very valuable resource to the field of transcription at large.

The main weakness of this paper is that the central conclusions are not convincingly supported by the data, as explained below.

1. Upon removal of Oct4, the authors found that some regions bound by Oct4 decrease in accessibility and some do not. However, the fact that some Oct4-bound regions do not require Oct4 to maintain their accessibility does not imply that Oct4 does not play a central role in regulating chromatin accessibility at other regions.<br /> Also note that regions bound by Oct4 but differentially dependent on Oct4 for their accessibility were described before using the same cell line (King and Klose, eLife 2017, Friman et al., eLife 2019).

2. Upon removal of Oct4, the authors found that regions maintaining their accessibility maintain Sox2 binding, while regions losing accessibility lose Sox2 binding. The authors use these findings (also already described before in the refs cited above) in support of a model where Sox2 transiently maintains accessibility in the absence of Oct4. The authors do not explain why Sox2 has a differential ability to maintain its binding in these two classes of regions. No Sox2 loss of function experiments were attempted to substantiate this statement.<br /> Friman et al., eLife 2019 defined regions that depend on Oct4, Sox2 or both of them for maintenance of their accessibility using the same Oct4 Tet-off cell line, as well as a Sox2 Tet-off cell line. Le Xiong et al. did not compare this dataset to theirs nor discuss it. Importantly, upon rapid Sox2 depletion, Friman et al. showed that more than half of Oct4 binding sites retained their accessibility. Thus, functional analysis has shown that Oct4 can maintain accessibility at a large fraction of its targets in the absence of Sox2.<br /> Taken together, their data together with previous literature converge on a different model, i.e. Oct4 controls chromatin accessibility at a (large) subset of regions it binds, and thereby regulates Sox2 binding. This explains why upon Oct4 knockdown, Sox2 binding decreases at regions that lose accessibility. In contrast, at regions bound by Oct4 but independent of Oct4 for their accessibility, Sox2 binding is maintained because chromatin accessibility does not change.

3. King et al., eLife 2017 have shown that Oct4 directly recruits the Brg1 subunit from the BAF complex, which colocalizes strongly with Oct4-bound regions in ES cells. This strongly suggests a direct role for Oct4 in the regulation of chromatin accessibility.

4. Friman et al., eLife 2019, performed rapid depletion of Oct4 using an Auxin-inducible system, and they observed a loss of accessibility at a large number of Oct4-bound regions that is quasi-synchronized with Oct4 loss. This also argues that Oct4 directly regulates chromatin accessibility.

5. Using the Tet-off Oct4 cell line, the authors observed a delayed loss of chromatin accessibility as compared to changes in transcriptional activity. From this observation, they conclude that Oct4 is a not crucial for regulating chromatin accessibility at these loci.<br /> However, this inference can only be true if there is an identical concentration-dependent activity of Oct4 in transcriptional activation and pioneer activity. Importantly, there is no reason to assume that this is the case. Transcriptional activity changes in response to changes in Oct4 levels might be very sensitive to slight decreases in Oct4 levels. Chromatin accessibility as observed by ATAC-seq might only start to decrease once Oct4 levels go below a certain threshold. In fact, it was reported (Strebinger et al., Molecular Systems Biology 2019) that cells with low endogenous Oct4 levels do not show changes in chromatin accessibility at pluripotency enhancers. This suggests that chromatin accessibility is relatively resilient to mild changes in Oct4 concentrations, which is what occurs after 3 hours of dox treatment in the present study.

6. The conclusions on the minor role of Oct4 in regulating chromatin accessibility are also weakened by the absence of Oct4 recovery experiments (i.e. dox treatment for 15-24 hours, and dox removal to re-express Oct4). In fact, Auxin-inducible degradation followed by recovery of Oct4 levels as well as recovery of Oct4 levels after mitotic degradation have shown to allow partial recovery in chromatin accessibility at a large number of Oct4-bound regions (Friman et al., eLife 2019). This also suggests a direct role for Oct4 in opening chromatin.

In summary, the data described in this paper are definitely very valuable. Their results allow to quantitatively describe the differential timing/sensitivity of transcriptional changes vs accessibility changes upon Oct4 knockdown, which is clearly new and insightful to understand the interplay between different mechanisms by which transcription factors regulate gene expression. A re-interpretation of this data could thus make this manuscript even more interesting.

Reviewer #1 (Public Review):

The authors aimed to develop a new, non-toxic tool for temporal regulation of Gal4-dependent gene expression in Drosophila, by creating a version of the Gal4 inhibitor, Gal80, bearing an auxin-degron sequence, rendering this protein susceptible to degradation upon provision of the plant hormone auxin by feeding. This technology (Auxin-inducible Gene Expression System, AGES) builds upon previous use of this system in other animals, including one study in Drosophila in which a different protein was targeted for auxin-dependent degradation (Trost, Fly 2016).

Strengths:

The authors have identified a need for a better tool for temporal control of transgene expression that is compatible with the vast libraries of Gal4 drivers, that doesn't rely on temperature shifts (as for Gal80ts), and that is non-toxic. As presented, they have been successful in developing such a tool and providing an initial characterization revealing its functionality, and key technical information for future exploitation (e.g. auxin dose, lag time of gene induction after auxin provision, the ability of auxin to cross the blood-brain-barrier).

Weaknesses:

1. The authors fail to give much credit to the previous work (Trost, Fly 2016), which provided the first demonstration of the utility, temporal dynamics, and non-toxicity of the auxin-degron system in Drosophila. While the current study applies the auxin-degron to generate a much more generally useful genetic tool, it is a bit ungenerous to only mention the early work in passing in the Introduction.

2. The technical testing of the system feels rather light for a tool-development manuscript, using AGES with two broadly expressed (and presumably quite strong) Gal4 drivers and a UAS-GFP effector transgene as a read-out of gene expression. Several simple extensions to this work would have been desirable in this first study.

For example:

- quantitative read-out of auxin-dependent GFP expression is only shown in Figure 2. Figures 3 and 4 show only images of a single animal in a given test condition. Such experiments could be quantified to give a sense of the animal-to-animal variation.

- the temporal dynamics of the system are only superficially described, despite the importance of this property for researchers to be aware of. The authors write (line 115-116): ""shorter exposure times of adults to auxin containing food were tested (data are not shown), however, 24 hours is the minimal amount of time required for proper induction of GAL4 activity.", but this is exactly the sort of information that should be shown and rigorously explored when presenting a new tool. Ideally, one could compare such properties side-by-side with Gal80ts. In addition, there is no mention of the reversibility of AGES (as is possible with Gal80ts), raising the question of how long auxin remains in the fly after ingestion.

- the authors argue the auxin provision is non-toxic, but the main read-out is survival/lifespan. While these are not affected by continuous exposure to auxin, the developmental time to pupal stages is clearly affected by high doses of auxin, so there is some pharmacological effect of this hormone. As such, more subtle effects of auxin (e.g., on locomotor activity, sleep, fertility etc.) cannot be fully excluded.

- the authors also write (line 205-6): "In our experience, auxin-containing food can be stored 4C for up to 4 weeks where the hormone's potency still persists", but, again, such observations would be much more useful to carefully document in this technical study.

The AGES system has the potential for use in the Drosophila community as a complementary and very valuable tool for temporal control of Gal4-driven gene expression. As with all tools, only time will tell whether the favorable properties highlighted by the authors' initial tests stand further scrutiny using other Gal4 drivers, other types of phenotypic read-out (gene expression, physiology, behavior etc).

Reviewer #2 (Public Review):

This paper's authors have developed a novel system that enables temporal control of Gal4-induced expression in Drosophila melanogaster. There are various drawbacks to existing systems used to control Gal4 expression, including toxicity and the need to shift temperatures, both of which are circumvented with the new AGES system described in this paper. The authors successfully display the efficacy of using this system to induce Gal4-driven reporter gene expression in the adult and larvae while also highlighting that the system is non-toxic and affordable and can be used in combination with existing Gal4 lines and will thus be of broad interest to the Drosophila community.

Reviewer #1 (Public Review):

This article focuses on a quantitative description of airineme morphology and its consequences for contact and communication between cells via these long narrow projections. The primary conclusions are

1) Airineme shapes are consistent with a persistent random walk model (analogous to a wormlike polymer chain), unhindered by the presence of other cells.

The authors convincingly demonstrate, using analysis of the mean-squared-displacement along the airineme contour, that the structures cannot be described by a diffusive growth process (ie: a Gaussian chain) as would be expected if there were no directional correlations between consecutive steps. Furthermore, by observing the airineme growth and looking at the distribution of step-sizes, they show that these steps do not exhibit the expected long-tail distributions that would imply a Levy-walk behavior. The persistent random walk (PRW) is presented as an alternative that is not inconsistent with the data. However, given the high level of noise due to low sampling, the claimed scaling behavior of the MSD at long lengths is not fully convincing. Nevertheless, the PRW provides a plausible potential description of the airineme shapes.

2) The flexibility (ie: persistence length) of the airineme shapes is one that maximizes the probability of a given airineme (of fixed length) contacting the target cell.

This optimum arises due to the balance between straight-line paths that reach far from the source but cover a narrow region of space and diffusive paths that compactly explore space but do not reach far from the starting point. Such optimization has previously been noted in unrelated contexts both for search processes of moving particles and for semiflexible chains that need to contact a target. The authors present a compelling case (Fig 4B) that the measured angular diffusion of the airinemes falls close to the predicted optimum. Furthermore, the measured probability of hitting the target cell also lies close to the model prediction, providing a strong test of the applicability of their model.

3) Airineme flexibility engenders a tradeoff between contact probability and directional information (ie: the extent to which the target cell can determine the position of the source).

This calculation proposes an alternative utility metric for communication via airinemes. The observed flexiblity is shown to be at a Pareto optimum, where changes in either direction would decrease either the probability of contact or the directional information. Again the absolute value of the metric (Fisher information for angular distribution) is within the predicted order of magnitude from the model. Thus, while the importance of maximizing this metric remains speculative, its quantitative value provides an additional test for the applicability of the PRW model.

Overall, this paper provides an interesting exploration of optimization problems for communication by long thin projections. A particular strength is the quantitative match to experimental data -- indicating not just that the experimental parameters fall along a putative optimum but also that the metrics being optimized are well-predicted by the model. Defining an optimization problem and showing that some parameter sits at the optimum is a common approach to generating insight in biophysical modeling, albeit invariably suffering from the fact that it is difficult to know which optimization criteria actually matter in a particular cellular system. The authors do an excellent job of exploring multiple optimization criteria, quantifying the balance between them, and pointing out inherent limitations in knowing which is most relevant.

A minor weakness of the manuscript is its focus on a very narrowly defined cellular system, with the general applicability of the results not being highlighted for clarity. For example, the fact that the same flexiblity optimizes contact probability and the balance between contact and directional information is an interesting conclusion of the paper. Is this true in general? Is it applicable to other systems involving a semiflexible structure reaching for a target or a moving agent executing a PRW?

Reviewer #2 (Public Review):

Signalling filopodia are essential in disseminating chemical signals in development and tissue homeostasis. These signalling filopodia can be defined as nanotubes, cytonemes, or the recently discovered airinemes. Airinemes are protrusions established between pigment cells due to the help of macrophages. Macrophages take up a small vesicle from one pigment cell and carry it over to the neighbouring pigment cell to induce signalling. However, the vesicle maintains contact with the source cell due to a thin protrusion - the airineme. In support of these data, the authors find that the extension progress of the airinemes fits an "unobstructed persistent random walk model" as described for other macrophages or neutrophils.

The authors describe the characteristics of an airineme as it would be a signalling filopodia, e.g. a nanotube or a cytoneme, which sends out to target a cell. An airineme, however, is fundamentally different. Here, a macrophage approaches a pigment cell binds to the airineme vesicle. Then, the macrophage approaches a target pigment cell and hands over the airineme vesicle. During this process, the airineme vesicle maintains a connection to the source pigment cell by a thin protrusion. Then, the macrophage leaves the target cell, but the airineme vesicle, including the protrusion, is stabilized at the surface and activates signalling. Indeed nearly all airinemes observed have been associated with macrophages (Eom et al., 2017).

Therefore, it is essential to focus on the "search-and-find" walk of the macrophage and not the passively dragged airineme. In the light of this discussion, I am not sure if statements like "allow the airineme to hit the target cell" are helpful as it would point towards an actively expanding protrusion like a filopodium.

Reviewer #3 (Public Review):

This paper studies statistical aspects of the role of long-range cellular protrusions called airinemes as means of intracellular communication. The mean square distance of an airineme tip is found to follow a persistent random walk with a given velocity and angular diffusion. It is argues that this distribution with these parameters is the one that optimise the probability of contact with the target cell. The authors then evaluate the directional information (where in space did the airineme come from) and found that, again, the measure diffusion coefficient optimise the trade-off between high directional information (small diffusion) and large encounter probability.

I found this paper well written and clear, and addressing an interesting problem (long-range intracellular communication) using rigorous quantitative tools. This is a very useful approach, which appears to have been appropriately done, that in itself makes this paper worthy of interest.

1. The main conclusion of this paper is that the airineme properties optimises something that has to do with their function. Although rather appealing, I find this kind of conclusion often questionable considering the large uncertainty surrounding many parameters. Here, optimality is shown from a practical perspective, using measure parameters. For instance, the optimal diffusion coefficient for hitting the target varies by 2 orders of magnitude when the distance between cells is varied (Fig.3A). The measured coefficient is optimal for cells about 25 µm distant. Does this reflect anything about the physiological situation in which these airinemes operate? Another rather puzzling claim is that the diffusion coefficient is optimised both for finding the target, AND for finding the best compromised between finding the target and providing directional information, while the latter must necessarily require weaker diffusion. Hence the last paragraph of p.6 ("the data is consistent with either conclusion that the curvature is optimized for search, or it is optimized to balance search and directional information"), although quite honest, gives the feeling that the conclusions are not very robust. I would welcome a discussion of these points.

2. on p.4: "the airineme tips (which are transported by macrophages [30]) appear unrestricted in their motion". I don't understand what it means that the airineme tips are transported by macrophage, and I missed the explanation in the cited article. Is airineme dynamics internally generated (i.e. by actin/microtubule polymerisation) or does it reflect to motility of cells dragging the airineme along? This is discussed in passing in the Discussion, but I think that this should be explainde in more detail right from the start. Aslo, if a cell is indeed directing the tip, what does contact mean? Does it mean that the driving macrophage must contact the target cell and somehow attached the airineme to it? IF yes, that means that the airineme tip has a large spatial extent, which will certainly affect the contact probability.

3. Fig. 2A shows the airinemes MSD and the fit using the PRW model. I don't find the agreement so good. The power law t^2 seems good almost up to 10 minutes, and the scaling above that, if there is one, is clearly larger than linear. So I would say that the apparent agreement with the PRW model reflects the fact that there is a crossover from a ballistic motion to something else, but that this something else is not a randow walk. The MSD does look quite strange at long time, where it apparently decays. This made me wonder whether there might be a statistical biais in the data, for instance, the longest living airinemes are those who didn't find their target and hence those who travel less far, on average. I tried to get more information on the data from the ref.[29,30], but could not find anything. The authors should discuss these data and possible biais in more detail. For instance, do the data mix successful and unsuccessful airinemes? This is somewhat touched upon in Fig.s$, but I did not gain any useful information from it, except that the authors find the agreement "good" while it does not look so good to me.

4. Regarding the directionality discussion, some aspect are a bit vague so that we are left to guess the assumptions made. For instance, the source cell is place at \theta=0 "without loss of generality" (p.6). Apparently (sketch Fig.5A) this also means that the airineme starting point from the source is at \theta=0, which clearly involves loss of generality, since the airineme could start from anywhere, its path could be hindered by the body of the source cell, and its contact angle would then be much less likely to be close to 0. It might be that in practice, only those airineme starting close to theta=0 do in fact make contact, but this should be discussed more thoroughly. Also, why is there to maxima in the Fisher information (Fig.5C) for very high and very low diffusion coefficient at short distance?

Reviewer #4 (Public Review):

In this work, the authors analyze the ability of zebrafish cells to find neighboring cells (playing the role of targets) by extended protrusions (airinemes) that display a "persistent random walk". In this kind of motion, a key parameter is the angular diffusivity, and the main finding is that, for the parameters of the experiments, the probability to find the target is maximized for the experimentally observed value of the angular diffusivity. Basically, a low value of angular diffusivity means that airinemes will be completely straight and miss easily small targets, but large values of angular diffusivities imply that the maximal spatial extension is reduced, disabling the cell to find targets that are to far away (as soon as one assumes that there is a maximal length of airinemes, as seems to be the case experimentally). As the result of a trade-off between these effects, there is an optimal value of angular diffusivity which turn out to be the experimentally observed value, for the experiments analyzed in the paper. In a second part, the authors ask whether the cell can gain directional information on where the target is located ; again the observed value of diffusivity seems to correspond to a trade-off between directional sensing and the ability to find targets.

I find that these results interesting since the paper brings arguments to understand which kind of stochastic motion are the most suitable to ease communication between cells. The analysis is performed by using simulations of persistent random walks, and image analysis of real cells. The main weakness of the approach consists in the fact that it is difficult, after having read the manuscript, to understand if the observed « optimality » seems is specific to the values of the cell sizes and cell-to-cell distances of the present experiment, or if it would also be approximately the case for other cell sizes, densities, etc. Nevertheless, quantifying such optimality for one experimental situation is already an interesting result.

Reviewer #1 (Public Review):

In this manuscript, the authors identified HOE-1, a tRNA processing enzyme, as an important regulator of UPRmt. They showed that nuclear HOE-1 is necessary and sufficient to activate UPRmt, which acts through ATFS-1 and DVE-1. The authors provided evidence that the UPRmt induced by nuclear retention of HOE-1 requires 3'-tRNA processing and tRNA transport. Moreover, HOE-1 is negatively regulated by ATFS-1 when UPRmt is activated. The experiments were well executed, and data are clear and convincing.

Comments:

1. The authors showed that HOE-1 localized to both mitochondria and nucleus in germline, while HOE-1(ΔNES) induces UPRmt in the intestine. The HOE-1 localization in the intestine should be presented including mitochondria and nucleus. The authors suggested that HOE-1 activates UPRmt in the intestine in a cell-autonomous manner. This would need to be demonstrated experimentally.

2. Whether 3'-tRNA processing is elevated in HOE-1(ΔNES) should be tested more directly. Is it possible to do determine the tRNA species that are elevated in HOE-1(ΔNES) strain by sequencing? Or that the authors can express HOE-1(ΔNES) that lacks the enzymatic activity and see whether it can still activate UPRmt.

3. The images shown in Fig. 8a is not clear. Enlarged images would be needed to clearly show changes in HOE-1 subcellular localization (mitochondria and nucleus) upon multiple mitochondria stresses.

4. The elevation of HOE-1 protein level is not clear (Fig. 8c, d). It is unclear whether the HOE-1 level in nuo-6 with atfs-1RNAi issignificantly increased compared with control RNAi or that in wild type with atfs-1 RNAi. The HOE-1 intensity in mitochondria vs nucleus would need to be examined in multiple mito stress conditions. It is also unclear how HOE-1 senses Mito stress.

Reviewer #2 (Public Review):

This manuscript reports that the tRNA processing enzyme HOE-1 is required for the activation of UPRmt in C. elegans. Given the dual-localization of HOE-1, the authors create mitochondrial and nuclear compartment-specific knockout of HOE-1 and demonstrated that only the nuclear HOE-1 is necessary and sufficient to activate the UPRmt. This paper will be of interest to scientists within mitochondrial stress response signaling. This study extends our understanding of how the mito-nuclear communication is mediated via the tRNA processing enzyme. However, some key aspects of the study need to be reinforced in the conclusions.

1. The phenotype that HOE-1(NLS) mutants suppressed the induction of the UPRmt in mitochondrial mutants is interesting. However, the mechanism underlying the HOE-1-mediated mitochondrial stress response is still not very clear. I have concerns regarding the specific involvement of HOE-1 in the regulation of UPRmt, since tRNA processing, the tRNA exporter xpo-3, as well as the RNase P complex popl-1, are all general regulators for protein synthesis. It is unclear how one can explain the specific involvement of these regulators only in the regulation of the UPRmt.

2. It is also confusing that HOE-1(NLS) mutants suppressed the UPRmt induction in nuo-6 mutants, however, xpo-3 which functions in the same pathway as HOE-1 in terms of tRNA processing and export did not suppress the UPRmt induction in nuo-6 mutants in Fig 6i and 6j.

3. The authors mentioned that HOE-1 homolog ELAC2 is not only required for tRNA maturation but also essential for the formation of tRNA fragments, snoRNAs, and miRNAs, are these non-coding RNAs account for the activation of the UPRmt?

4. It is interesting to show that hoe-1(ΔNES) mutant is sufficient to induce the nuclear accumulation of the ATFS-1 and the subsequent up-regulation of the UPRmt reporter gene. However, the authors did not rule out the possibility that mitochondrial protein homeostasis was already disrupted in hoe-1(ΔNES) mutants so that the UPRmt was induced.

5. The authors only showed that mitochondrial membrane potential was not changed in hoe-1(ΔNES) mutants. More characterization of mitochondrial function in hoe-1(ΔNES) mutants is required, such as OCR and mitochondrial morphology. It seems that hoe-1(ΔNES) mutants are smaller than wild-type animals.

6. In fig 4a, Why the overall level of ATFS-1 is dramatically increased in hoe-1(ΔNES) mutants, this is not consistent with only two-fold up-regulation of atfs-1 transcript levels. The authors also would need to show the ATFS-1::GFP expression pattern in the nuo-6 mutants as a control.

Reviewer #3 (Public Review):

Held and colleagues present numerous intriguing findings suggesting that the tRNA processing enzyme ELAC2/HOE-1 is required to activate the mtUPR in C. elegans. The hoe-1 gene encodes 2 proteins, one of which contains a mitochondrial targeting sequence (MTS) and a nuclear localization sequence (NLS). The other protein is similar but lacks the MTS. Thus, hoe-1 encodes proteins involved in tRNA processing in the nucleus and within mitochondria. It is intriguing that one or both of the proteins may be required for mtUPR activation. I have multiple concerns related to the experimental design and the interpretation. However, my major concern is that it remains unclear how HOE-1 regulates the mtUPR (DVE-1 or ATFS-1).

Major concerns.

-Figure 1. The authors use the transcriptional reporter hsp-6::gfp as a mtUPR reporter. However, in addition to requiring transcription for the hsp-6 promoter to induce the gfp mRNA, that mRNA must be synthesized. As HOE-1 is a tRNA processing enzyme likely required for protein synthesis, qRT-PCR analysis should be performed to quantify the effects of HOE-1 inhibition on the mtUPR transcription response. Thus the data supporting the claim that loss-of-function mutations in HOE-1 inhibit mtUPR-dependent transcription are weak and must be further substantiated.

- Several groups have shown that inhibition of S6 kinase inhibits mtUPR activation. As HOE-1 is presumably required for protein synthesis, perhaps the mechanism is related? It would be good to know whether the inhibition of other genes affecting tRNA levels also impairs mtUPR or is specific to HOE-1.

-It is my understanding that the HOE-1 protein with a mitochondrial targeting sequence is transcribed from the same gene as HOE-1 without the MTS. And, there are separate transcriptional start sites for each mRNA/protein. Considering the number of claims related to subcellular localization of HOE-1, the authors would need to determine if transcription from either site is altered during mitochondrial stress.

- There is an over-reliance on the hoe-1(∆NES) strain which causes mtUPR activation. It remains unclear if nuclear accumulation is an event driving mtUPR activation or if the activation is simply an artifact of the ∆NES mutation. The hoe-1 loss of function studies are need to be further developed in order to interpret the hoe-1(∆NES) results. It remains possible that the ∆NES findings are simply an artifact of a neomorphic allele and do not inform on HOE-1 function.

-The data suggesting that nuclear accumulation of HOE-1 is sufficient to activate mtUPR is relatively weak. Does HOE-1∆NES cause mitochondrial dysfunction which increases mtUPR activation? Potentially, HOE-1 lacking the nuclear export sequence may not accumulate within mitochondria and cause mitochondrial dysfunction. More in depth quantitative assessment of mitochondrial activity is required (TMRE images, oxygen consumption, etc). Alternatively, the ∆NES mutation could be combined with the ∆MTS mutation.

-Fig 4. The authors generate a beautiful ATFS-1::mCherry fusion protein and demonstrate that accumulates within nuclei during mitochondrial stress. Does hoe-1 inhibition affect translation/synthesis of ATFS-1::mCherry or nuclear accumulation of ATFS-1::mCherry? Or, DVE-1?

The mechanism by which hoe-1 impacts mtUPR is unclear.

Reviewer #1 (Public Review):

The study visualizes the behavior of some of the components of a cell stress pathway in live cells. The study generates tools that may be of interest to cell biologists, though the claims of the study need to be tempered to better reflect what is actually observed and some of the reagents would benefit from additional characterization.

The authors have applied some live cell imaging tools to attempt to visualize the processing of XBP1 mRNA by IRE1a during the mammalian Unfolded Protein Response. Single particle tracking was combined with the MS2 tagging system to localize wt and mutant XBP1 mRNAs relative to the endoplasmic reticulum (ER). This is the first study to visualize XBP1 mRNA in a live cell and the information acquired supports existing models of XBP1 mRNA processing and potentially provides some clarity regarding spatial localization and rates of processing in live cells. The manuscript makes some claims that need to be modified as the data are sometimes more limited in terms of what is actually shown. In addition, the authors perform some live cell imaging experiments with a tagged version of IRE1a, the stress sensor that cleaves XBP1 mRNA as part of its splicing process during stress. Previous studies have reported that IRE1a forms large visible clusters in response to ER stress. The authors have claimed that the clustering is an artifact of tagged IRE1a overexpression. More characterization of both the reporter and native untagged IRE1a are needed to make a stronger conclusion. Overall, the study will be of interest to labs in the ER stress field and of potentially broader interest to groups studying mRNA trafficking and processing in live cells. With further characterization, the reagents may be useful for mechanistic studies of ER stress in single cells.

Reviewer #2 (Public Review):

This manuscript develops different reporters to monitor XBP1 targeting to the ER, which are used to confirm previous results showing that XBP1 is directed to the ER through a mechanism involving translation of the HR2 mRNA sequence. As indicated in the manuscript, this mechanism had been previously reported by Kohno, and, while the work presented here confirms this model, it does not extend it. The major advance from this manuscript, apart from the reporter development, relates to the fact that IRE1 clusters are not observed in cells expressing endogenous levels of IRE1-GFP and subjected to ER stress. This is in contrast to previous reports where IRE1 clusters were proposed to be the primary site of XBP1 splicing; however, IRE1 clustering from XBP1s splicing has been shown to been separable previously in Ricci et al (2019) FASEB J (where they showed that the flavinoid luteolin induces robust XBP1 splicing independent of clustering). Herein, the authors demonstrated that the clustering of IRE1-GFP is an artifact of overexpression, which is not observed upon expression of IRE1-GFP to endogenous levels.

Ultimately, while the experiments appear well performed, the advance of this current manuscript is limited. The data included in Fig. 1-3 validate previous mechanisms proposed for XBP1 targeting to the ER using new approaches. While important to validate mechanisms using different approaches, there is no new insight included in this aspect of the work. The fact that IRE1 clustering results from an artifact resulting from overexpression of IRE1-GFP is important, although it is somewhat underdeveloped in this specific manuscript. However, this report does support findings in a recent preprint posted to bioRXIV (Belyy et al (2021)), similarly showing that IRE1 does not cluster, as previously, thought. Taken together, this work and the Belyy et al preprint does indicate that IRE1 clustering is not associated with activation, but instead represents an artifact of overexpression.

Reviewer #3 (Public Review):

This manuscript applies single molecule imaging approaches to visualize the ER targeting of Xbp1 mRNA by the unfolded protein response and its processing by IRE1. The major conclusions are that translation of the hydrophobic HR2 domain localizes a portion of Xbp1 mRNA to the ER, that ER stress releases Xbp1 from the ER due to splicing action by IRE1, and that Xbp1 mRNA appears not to make stable associations with punctal clusters of IRE1 during ER stress, and that these clusters do not appear in this cell system at lower levels of ectopic IRE1 expression, potentially calling the role of these clusters in Xbp1 splicing. The strength of the work is in using single molecule imaging to test (and largely confirm) ideas that were previously advanced in the literature based on studies of lower resolution, although the apparent lack of a functional role for IRE1 clusters at least in this system addresses a point that remains unsettled in the field.

The authors take advantage of tandem tagging to label both Xbp1 mRNA and the polypeptide product associated with translation of unspliced Xbp1 mRNA. The authors show convincingly that fluorescent dots corresponding to unspliced Xbp1 associate with the ER, to an extent greater than that achieved by Xbp1 in which the HR2 peptide cannot be translated, but to an extent less than that achieved by a conventional secretory protein. Why the Xbp1 mRNA achieves lower targeting efficiency is not specified. They also show that ER stress is associate with a loss of Xbp1 mRNA from the ER, and this is attributable to splicing by Xbp1, whereas unspliceable Xbp1 or wild-type Xbp1 when IRE1 is inhibited remains associated with the ER to an extent during stress that is not much lower than in the absence of stress. Conversely, constitutively spliced Xbp1 largely fails to associate with the ER. The last figure leverages the imaging of Xbp1 to show that Xbp1 mRNA appears not to associate with IRE1 clusters that are observed in a system similar (but not identical) to the cell system reported by the Walter lab in 2020.<br /> Overall, the experiments are intriguing and the quality of the data is high. The major novelty of the paper is its approach. The general findings (that the HR2 region of Xbp1 mRNA must be translated for Xbp1 mRNA to be targeted to the ER; and that splicing of Xbp1 mRNA, which shifts the reading frame of the HR2 region, causes Xbp1 mRNA to no longer be associated with the ER largely support/confirm the conclusions of the field arrived at through other methods. The last conclusion, about IRE1 clustering, is where new ground is tread. It is notable that IRE1 clusters are not observed when IRE1 is ectopically expressed to low levels. Therefore, phenomenologically at least, IRE1 clusters are not a prerequisite for at least some splicing of Xbp1 mRNA to occur.

All that said, I have four substantive concerns about the manuscript:

1. The conclusions with respect to Xbp1 mRNA (and also with respect to XBP1 translation) require that the visualized Xbp1 dots are indeed single molecules of Xbp1 mRNA, and that the process captures all Xbp1 mRNA molecules, rather than only a subpopulation. I am not so sure that either of these criteria is rigorously validated. In Supplementary Figure 2, it appears that MCP-Halo and scAB-GFP detect many spots that are either overlapping or immediately adjacent to each other - more than I would expect by chance given that these two sorts of spots arise necessarily from different RNAs. This raises the possibility that what is detected are not individual RNAs but clusters thereof. If that is true (or even if it isn't), there also needs to be some way of validating that the technique is not biasing for only a certain population of Xbp1 mRNAs that behave in a certain way that is not necessarily representative of all Xbp1 mRNAs. Indeed, the fact that FISH detects some Xbp1 mRNAs that scAB-GFP does not (Figure 2F) argues that scAB-GFP is not detecting everything, which raises the question of what features characterize mRNAs that it does not detect.

2. I agree with the authors' interpretation that Xbp1 mRNA (or at least the Xbp1 mRNA that is being detected) does not stably associate with IRE1 clusters. However, it is not clear that one would expect a stable association. Rather, is it not possible that splicing might be by a "kiss-and-run" mechanism? To test/eliminate this possibility, the authors would need to show the fate of individual Xbp1 mRNAs before and after an IRE1 encounter and/or before and after leaving the ER. It would seem that the authors have the tools to accomplish this in their existing toolkit.

3. The conclusion that splicing of Xbp1 mRNA causes its liberation from the ER membrane is largely inferential. I agree it is a reasonable conclusion, but, similarly to point 2 above, it requires tracking the mRNA before and after its cleavage and/or before and after its release from the ER to conclusively validate.

Reviewer #1 (Public Review):

In this detailed study the authors show that in isolated islets the polarity of the secretory apparatus is largely lost while it is preserved in slices where the capillary network remains intact. The authors then go on to show that the integrin/FAK pathway appears to be responsible for inducing and maintaining polarity, which involves concentration of active zone proteins and calcium channels at the contact sites and a higher sensitivity and potency of insulin secretion to glucose stimulation.

Generally, the data appear to be of high quality, being carried out with state-of-the-art technology, and the manuscript is lavishly illustrated. Since as a neuroscientist I am not sufficiently familiar with the field of the cell biology of insulin release it is difficult for me to judge whether there is sufficient advance in knowledge. A higher degree of organization of release sites including a role of active zone proteins was previously demonstrated from other endocrine organs involving the release of large dense-core vesicles such as chromaffin cells. Thus, the differences between the highly organized and rapidly responding exocytotic sites in neurons and the slower reacting release sites of peptide/protein containing granules are not fundamental but rather gradual, despite the principal cell biological differences between the biogenesis and recycling pathways of the secretory organelles.

In summary, the work adds new aspects to the understanding of the regulation of exocytosis in pancreatic beta cells. Aside from corrections of figure descriptions and experimental details, my only major comment relates to the data shown in Fig. 4. It appears that the difference in the time-to-peak between the two preparation is mainly caused by a (rather variable?) delay between glucose addition and the onset of the rise since the rate of increase is apparently not different between the preparations. Is this due a delay in depolarization, i.e. a delay in the closure of the ATP-K channels? This should be clarified. Also, the authors should show a comparative histogram of the delay times (between glucose addition and the inflection point at the onset of the rise).

Reviewer #2 (Public Review):

1. The authors present an investigation of subcellular distribution and dynamics of known presynaptic proteins in a relatively new approach, pancreatic slices, mastered by a limited number of laboratories, and which is currently the best method to largely preserve capillary networks. They demonstrate the advantage of this method by detailed cellular and subcellular optical analysis comparing isolated islets, islets in pancreatic slices, isolated islet cells and isolated islet cells on ECM (laminin) covered surfaces. This work provides good proof that preservation of capillary networks and corresponding distribution of proteins (laminin, liprin, integrin beta1 etc) is required for insulin secretion at the apical surface of islet cells. Moreover, in these pancreatic slices they observe a restriction of exocytotic sites at the vascular surfaces. The role of the extracellular matrix is also well investigated here by experiments on dispersed or single beta cells attached either to a glass-BSA interface or to a glass-laminin interface.<br /> However, the authors have already previously published in 2014 a restricted polarized insulin secretion in cultured islets as well as the preservation of localized liprin and laminin distribution (as well as RIM2 and piccolo; DOI 10.1007/s00125-014-3252-6). It is not clear why these data cannot be reproduced now again in isolated islets (see Fig. 1 and 2) .

2. The authors try to gain insight which mechanisms control this specific spatial restriction and they provide evidence that Focal Adhesion kinase activity is implicated in glucose-induced calcium fluxes and insulin secretion by the use of a small molecule antagonist and the use of a purified monoclonal antibody. They conclude that FAK is a master regulator of glucose induced insulin secretion that controls positioning of presynaptic scaffold proteins and the functioning of calcium channels.<br /> Although FAK may be a regulator, the claim that FAK controls functioning of calcium channels can certainly not be made. Ratio measurements of cellular calcium levels do not suffice for that (patch or sharp would be required). Moreover, the fact that KCl-induced insulin secretion (which bypasses nutrient metabolism and leads directly to opening of voltage-dependent calcium channels) is not altered by the FAK antagonist strongly argues against a role of FAK in calcium channel regulation. Indeed, the presented data suggest that FAK may intervene far more upstream from exocytosis such as in nutrient metabolism or granule mobility/maturation.

3. The authors present data that islets in pancreatic slices are considerably more sensitive to glucose, inducing a response already at basal glucose levels (2.8 mM). In the same vein the authors observe a considerably shortened delay between stimulus and response (this delay is general due to nutrient metabolism and initial filling of intracellular calcium stores). The authors take these phenomena as evidence for a superior and more physiological quality of their islet slices as compared to conventional purified islets.

However, contrary to their interpretation, these observations considerably questions whether the slice preparation used here in this work has physiological qualities. Indeed, the authors observe considerable activity of islet beta-cells already far below the set-point of around 6 or 7 mM in rodents, very well characterized through a number of studies in-vivo, in-vitro and even in-situ (10.1113/jphysiol.1995.sp020804), and their preparations reach almost full activity around the set-point. This is also surprising as such a hypersensitivity has not been reported by several other groups using the same preparation, i.e. pancreatic slices (10.1152/ajpendo.00043.2021; 10.1371/journal.pone.0054638; 10.3389/fphys.2019.00869; 10.1371/journal.pcbi.1009002; 10.1038/nprot.2014.195) even using patch clamp (10.3390/s151127393). Moreover, even human islets, known for a lower set-point, are inactive in slices at 3 mM (10.1038/s41467-020-17040-8) in line with the physiological requirement to avoid insulin secretion in low glucose states as to avoid life-threatening hypoglycaemia. The same applies for the shortened delay between application of a stimulus (glucose) and start of the response, which has also not been observed by other groups in pancreatic slices (refs see above).

In general, such an increased glucose sensitivity is observed in prediabetic states or experiments mimicking such a condition. To the best of my recollection such an apparently increased sensitivity can also be observed in brain slices due to leakage. Unfortunately, no independent measures of islet quality in slices are provided.<br /> Within the same vein the comparison between slices and islets (Fig 5) is not in favour of a more physiological aspect of slices and the different cell morphology and small number of observations shed more doubt, especially in view of the well known normal beta-cell heterogeneity (which may explain differences and may have been missed here due to a small sample size).

In a larger context this glucose supersensitivity may also shed doubts on the proposed important role of FAK as its role may be far less preponderant in preparations corresponding to physiological criteria.

Reviewer #3 (Public Review):

The authors have - using 2-photon imaging of exocytosis - compared insulin secretion in isolated islets maintained in tissue culture with that in islets of acutely prepared slices of the pancreas.

They demonstrate that i) secretion is highly polarized and direct towards the capillaries in the latter preparation, is ii) that the concentration dependence of insulin secretion is shifted towards lower concentrations; and iii) that the difference involves presynaptic scaffold proteins and focal adhesion kinase (FAK).

The manuscript is well written with a clear narrative and the data are logically presented.

The experimental data are of high quality (typifying the work of this team), the results are unexpected but of great significance. The findings bridge the gap between in vivo and in vitro studies.

Insulin secretion becomes defective in type 2 diabetes (T2D; the commonest form of diabetes afflicting close to 400 million individuals worldwide) and if these data can be extended to human pancreases they might be of highly relevant to the understanding of the aetiology of T2

In general, I am enthusiastic about the paper. Possibly, the mechanisms underlying the differences between isolated islets and slices could be explored in somewhat greater detail and there a actually a few loose ends but they may be possible to resolved by textual changes and may not require additional work.

Reviewer #1 (Public Review):

Sato et al. investigated the role of the thalamus-derived factor, VGF, as extrinsic cue that controls layer 4 development in the cortex. They show that this process is necessary for further maturation of the primary somatosensory cortex (S1) and the barrel field formation. To explore the role of thalamocortical axon (TCA) projections in cortical layer formation, the authors developed a mouse model with TCA ablation from the ventrobasal (VB) thalamic nucleus by the administration of diphtheria toxin (DT) from P0. They induced a decrease in the VB nucleus size and the TCAs terminals were also diminished in layer 4 of S1. Sato et al. demonstrated that the number of layer 4 cells in S1 is reduced in TCA-ablated model and to verify that these TCA ablation from the VB was indeed responsible for this laminar phenotype, they used a different approach to specifically ablate only the VB neurons. By performing a technically impressive in utero electroporation at e11.5 in the thalamus with a diphtheria toxin receptor (DTR) expression plasmid and then administered DT at P0, they mimicked the laminar phenotype in the cortex and the results were similar to the TCA-ablated mice. Moreover, they showed that, apparently, the rest of the cortical layers remain intact. Interestingly, the authors demonstrated that VGF and NRN1 as TCA-derived factors play an important role to maintain the layer 4 neuronal number during cortical development by restoring these cell number after the overexpression of these factors in TCA-ablated model in vivo. To further address this question, they induced a genetic inactivation of Vgf by using CRISPR/Cas9-mediated gene editing, and they proved that VGF is necessary for the maintenance of layer 4 neurons number.

The manuscript shows potentially interesting findings but there are some open questions and experiments that should be done to better support the conclusions of the paper. Moreover, some aspects of image acquisition and data analysis need to be clarified.

Reviewer #2 (Public Review):

Sato H. and colleagues here investigate the role of extrinsic factors in the development of the murine neocortex. It was previously shown by these authors that thalamocortical neurons, through the expression of Vgf and Nrn1 among others, contribute specifically to layer 4 neurons development in vitro. In the current study, by postnatally ablating thalamocortical projections and studying a Vgf knockout, they further investigate the role of these projections in layer 4 establishment and bring some light to the in vivo role of VGF.

Although this is an interesting study, the novelty is relatively limited as it incrementally builds on previous work from this laboratory as well as previous work from several laboratories directly addressing the effects of ablation of specific thalamic nuclei on cortical neuron identity. In order to increase interest and relevance, several key experiments should be performed.

Reviewer #3 (Public Review):

The first extrinsic influences that shape the cortical neuroepithelium are secretory factors, emanating from signaling centers adjacent to the telencephalic vesicles. These centers set up the gross areal pattern of the neocortex without any extrinsic signals. While it is clear that there are intrinsic gradients from the beginning of cortical neurogenesis, there are also extrinsic cues that contribute to the differences. The best candidate to deliver the area-specific cues to the cortex is via area-specific thalamocortical projections. These arrive to the cortex very early, at the peak of the cortical neurogenesis and neuronal migration. The impact of thalamic lesions on cortical lamination was demonstrated by Windrem and Finlay, 1991. Moreover the influence of ephrin A5 on cortical progenitor cells and the effect of secreted Wnt3 on neuronal differentiation has been previously described.

The Sato et al., paper studies two thalamus-derived factors that might mediate some of the area-specific differences to the cortex. The authors used the results of their previous screens to identify secreted molecules that are not produced in the cortex, but delivered to the cortex through the thalamocortical projections. The authors screened for thalamus-specific genes by comparing expression profiles of the thalamus and the cortex (Sato et al., 2012). These screens identified neuritin 1 (NRN1) and VGF nerve growth factor inducible (Vgf) transported to the cortex through TCAs. This study demonstrates that VGF maintains the proper amount of layer 4 neurons in S1.

The study is reported in a logical sequence. First the authors established that birthdates of all cortical layers in various cortical areas are all prenatal. Since all neurons were generated before birth, no NeuN positive birthdated cells were found after postnatal EdU injection and therefore it is unlikely that layer 4 neurons are additively generated in S1. The study also analysed the emergence of RORβ (RAR-related orphan receptor beta) expression in the postnatal primary somatosensory cortex.

Then, they performed toxin-mediated ventrobasal complex ablation in vivo. Cre expression was the highest in VB among the thalamic nuclei in 5HTT-Cre mice, therefore the lesion was the greatest there. This reduced the thalamocortical axons that project to primary somatosensory cortex. The ablation triggered accumulation of Iba1-immunoreactive microglial cells in VB. This indicated that these are the regions with dead cells, in the thalamus. After this ablation VB was reduced at P5, and RORα-expressing thalamic neurons were decreased.

The authors examined the possibility that other thalamic neurons project to the S1, but detailed tracing from S1 did not produce backlabeling pattern in the thalamus that would indicate TCA ingrowth from the dLGN or MG to S1. No such re-wiring was observed in the PO VB TCA-ablated mice.

The number of RORβ-expressing cortical cells was decreased to 67% as compared with control cortex. Moreover the absolute number of layer 4 neurons also decreased in the TCA-ablated S1. This argued against the possibilities of altered RORβ expression or fate change of the layer 4-destined cells to those of other layer. The authors suggested cell death as a possible mechanisms for getting these differences. However, all the conventional cell death detection methods (ssDNA, cleaved caspase 3, Iba1, mRNA of Bax, Bad, and Bak, and DAPI), they could not obtain convincing evidence for significant cell death induction in layer 4 upon TCA ablation.

The toxin-mediated ventrobasal complex ablation in vivo is based on cre expression. Since Cre expression was the highest in VB among the thalamic nuclei in 5HTT-Cre mice, this was the region for the greatest damage. Nevertheless there was additional cre expression in layer 6 and also in the raphe nucleus the authors designed experiments to exclude that the layer 4 reduction in 5HTT-Cre; R26-DTR mice is due to ablation of these brain parts rather than the VB in the thalamus. The authors used DTR expression plasmid that was electroporated into the embryonic dorsal thalamus in utero at E11.5, when VB neurons are generated. The authors demonstrated that the effect was S1 specific. The cortical areas with dLGN and MGN innervation were not affected. Layers 2/3 (revealed with Brn2), layer 5 (revealed with Ctip2) and layer 6 (revealed with Tbr1) appear intact upon TCA elimination suggesting that the effect was specific for layer 4. The number of layer 4 neurons is restored by forced expression of NRN1 and VGF in the cortex of TCA-ablated mice NRN1 and VGF was lost in the thalamic nuclei and their axon terminals in layer 4 of S1 in TCA-ablated mice

The next group of experiments demonstrated that genetic inactivation of Vgf in thalamocortical projection neurons results in a reduction in layer 4 neurons in S1 cortex and that NRN1 is dispensable in this process. The authors used three single-guide RNAs (sgRNAs) cutting exons of Nrn1 and Vgf to induce frame shifts resulting in failure of protein translation of both NRN1 and VGF. They electroporated the sgRNAs and Cas9 protein into fertilized eggs, mutations were induced in the genomic sequences of Nrn1 and Vgf allele near designed sgRNAs. The loss of TCA-derived VGF from the cortex resulted in the significant reduction of RORβ immunoreactive layer 4 cells in S1 and in V1. These observations suggest that the regulation of the neuronal number of layer 4 by TCAs via VGF is a common mechanism operating widely in sensory areas.

VGF released from TCA terminals sets the exact numbers of cortical layer 4 neurons. Then, the activity dependent sorting of thalamocortical afferents will impose the cytoarchitectonic changes that will form the cytoarchitectonic barrels. This interaction between thalamic and layer 4 neurons to form the cytoarchitectonic barrel formation in S1 was significantly impaired in Vgf-KO mice despite the presence of TCAs. The paper convincingly demonstrates that thalamocortical axons play instructive roles in the regulation of layer 4 cell numbers and the specification of area properties of somatosensory and visual cortices.

I consider this study a very important and significant step in our field. This is a direct demonstration of the instructive role of the thalamocortical projections. The major conclusions of the study will have to be supported by additional experiments (cell death - fate change distinction).

Reviewer #1 (Public Review):

In this manuscript, Saxena, Russo et al. study the principles through which networks of interacting elements control rhythmic movements of different speeds. Typically, changes in speed cannot be achieved by temporally compressing or extending a fixed pattern of muscle activation, but require a complex pattern of changes in amplitude, phase, and duty cycle across many muscles. The authors train an artificial recurrent neural network (RNN) to predict muscle activity measured in monkeys performing an arm cycling task at different speeds. The dominant patterns of activity in the network do not directly reflect muscle activity. Instead, these patterns are smooth, elliptical, and robust to noise, and they shift continuously with speed. The authors then ask whether neural population activity recorded in motor cortex during the cycling task closely resembles muscle activity, or instead captures key features of the low-dimensional RNN dynamics. Firing rates of individual cortical neurons are better predicted by RNN than by muscle activity, and at the population level, cortical activity recapitulates the structure observed in the RNN: smooth ellipses that shift continuously with speed. The authors conclude that this common dynamical structure observed in the RNN and motor cortex may reflect a general solution to the problem of adjusting the speed of a complex rhythmic pattern. This study provides a compelling use of artificial networks to generate a hypothesis on neural population dynamics, then tests the hypothesis using neurophysiological data and modern analysis methods. The experiments are of high quality, the results are explained clearly, the conclusions are justified by the data, and the discussion is nuanced and helpful. I have several suggestions for improving the manuscript, described below.

1. It would be useful for the authors to elaborate further on the implications of the study for motor cortical function. For example, do the authors interpret the results as evidence that motor cortex acts more like a central pattern generator - that is, a neural circuit that transforms constant input into rhythmic output - and less like a low-level controller in this task? The observation that cortical activity looks more like the pattern-generating modes in the RNN than the EMG seem to be consistent with this interpretation. On the other hand, speed-dependent shifts for motor cortical activity in walking cats (where the pattern generator survives the removal of cortex and is known to be spinal) seems qualitatively similar to the speed modulation reported here, at least at the level of single neurons (e.g., Armstrong & Drew, J. Physiol. 1984; Beloozerova & Sirota, J. Physiol. 1993). More generally, the authors may wish to contextualize their work within the broader literature on mammalian central pattern generators. For instance, some conclusions of this study seem to parallel experimental work on the locomotor CPG, where a constant input (electrical or optogenetic stimulation of the MLR at a frequency well above the stepping rate) drives walking, and changes in this input smoothly modulate step frequency.


2. If the input to the RNN were rhythmic, the network dynamics would likely be qualitatively different. The use of a constant input is reasonable, but it would be useful for the authors to elaborate on this choice and its implications for network dynamics and control. For example, one might expect high tangling to present less of a problem for a periodically forced system than a time-invariant system. This issue is raised in line 210ff, but could be developed a bit further. The use of a constant input should also be discussed in the context of cortical physiology, as motor cortex will receive rhythmic (e.g., sensory) input during the task. The argument that time-varying input to cortex will itself be driven by cortical output (475ff) is plausible, but the underlying assumption that cortex is the principal controller for this movement should be spelled out. Furthermore, this argument would suggest that the RNN dynamics might reflect, in part, the dynamics of the arm itself, in addition to those of the brain regions discussed in line 462ff. This could be unpacked a bit in the Discussion.


3. The low tangling in the dominant dimensions of the RNN is interpreted as a signature of robust pattern generation in these dimensions (lines 207ff, 291). Presumably, dimensions related to muscle activity have higher tangling. If these muscle-related dimensions transform the smooth, rhythmic pattern into muscle activity, but are not involved in the generation of this smooth pattern, one might expect that recurrent dynamics are weaker in these muscle-related dimensions than in the first three principal components. That is, changes along the dominant, pattern-generating dimensions might have a strong influence on muscle-related dimensions, while changes along muscle-related dimensions have little impact on the dominant dimensions. Is this the case?


4. It would be useful to have more information on the global dynamics of the RNN; from the figures, it is difficult to determine the flow in principal component space far from the limit cycle. In Fig. 3E (right), perturbations are small (around half the distance to the limit cycle for the next speed); if the speed is set to eight, would trajectories initialized near the bottom of the panel converge to the red limit cycle? Visualization of the vector field on a grid covering the full plotting region in Fig. 3D-E with different speeds in different subpanels would provide a strong intuition for the global dynamics and how they change with speed.


5. What was the goodness-of-fit of the RNN model for individual muscles, and how was the mean-squared error for the EMG principal components normalized (line 138)? It would be useful to see predicted muscle activity in a similar format as the observed activity (Fig. 2D-F), ideally over two or three consecutive movement cycles. A related issue is whether the solutions are periodic for each individual node in the 50-dimensional network at each speed (as is the case for the first few RNN principal components and activity in individual cortical neurons and the muscles). If so, this would seem to guarantee that muscle decoding performance does not degrade over many movement cycles. Some additional plots or analysis might be helpful on this point: for example, a heatmap of all dimensions of v(t) for several consecutive cycles at the same speed, and recurrence plots for all nodes. Finally, does the period of the limit cycle in the dominant dimensions match the corresponding movement duration for each speed?


6. How does the network respond to continuous changes in input, particularly near zero? If a constant input of 0 is followed by a slowly ramping input from 0-1, does the solution look like a spring, as might be expected based on the individual solutions for each speed? Ramping inputs are mentioned in the Results (line 226) and Methods (line 805), but I was unable to find this in the figures. Does the network have a stable fixed point when the input is zero?


7. Why were separate networks trained for forward and backward rotations? Is it possible to train a network on movements in both directions with inputs of {-8, ..., 8} representing angular velocity? If not, the authors should discuss this limitation and its implications.


8. It is somewhat difficult to assess the stability of the limit cycle and speed of convergence from the plots in Fig. 3E.  A plot of the data in this figure as a time series, with sweeps from different initial conditions overlaid (and offset in time so trajectories are aligned once they're near the limit cycle), would aid visualization. Ideally, initial conditions much farther from the limit cycle (especially in the vertical direction) would be used, though this might require "cutting and pasting" the x-axis if convergence is slow. It might also be useful to know the eigenvalues of the linearized Poincaré map (choosing a specific phase of the movement) at the fixed point, if this is computationally feasible.

Reviewer #2 (Public Review):

The study from Saxena et al "Motor cortex activity across movement speeds is predicted by network-level strategies for generating muscle activity" expands on an exciting set of observations about neural population dynamics in monkey motor cortex during well trained, cyclical arm movements. Their key findings are that as movement speed varies, population dynamics maintain detangled trajectories through stacked ellipses in state space. The neural observations resemble those generated by in silico RNNs trained to generate muscle activity patterns measured during the same cycling movements produced by the monkeys, suggesting a population mechanism for maintaining continuity of movement across speeds. The manuscript was a pleasure to read and the data convincing and intriguing. I note below ideas on how I thought the study could be improved by better articulating assumptions behind interpretations, defense of the novelty, and implications could be improved, noting that the study is already strong and will be of general interest.

Primary concerns/suggestions:

1. Novelty: Several of the observations seem an incremental change from previously published conclusions. First, detangled neural trajectories and tangled muscle trajectories was a key conclusion of a previous study from Russo et al 2018. The current study emphasizes the same point with the minor addition of speed variance. Better argument of the novelty of the present conclusions is warranted. Second, the observations that motor cortical activity is heterogenous are not new. That single neuronal activity in motor cortex is well accounted for in RNNs as opposed to muscle-like command patterns or kinematic tuning was a key conclusion of Sussillo et al 2015 and has been expanded upon by numerous other studies, but is also emphasized here seemingly as a new result. Again, the study would benefit from the authors more clearly delineating the novel aspects of the observations presented here.

2. Technical constraints on conclusions: It would be nice for the authors to comment on whether the inherent differences in dimensionality between structures with single cell resolution (the brain) and structures with only summed population activity resolution (muscles) might contribute to the observed results of tangling in muscle state space and detangling in neural state spaces. Since whole muscle EMG activity is a readout of a higher dimensional control signals in the motor neurons, are results influenced by the lack of dimensional resolution at the muscle level compared to brain? Another way to put this might be, if the authors only had LFP data and motor neuron data, would the same effects be expected to be observed/ would they be observable? (Here I am assuming that dimensionality is approximately related to the number of recorded units * time unit and the nature of the recorded units and signals differs vastly as it does between neuronal populations (many neurons, spikes) and muscles (few muscles with compound electrical myogram signals). It would be impactful were the authors to address this potential confound by discussing it directly and speculating on whether detangling metrics in muscles might be higher if rather than whole muscle EMG, single motor unit recordings were made.

3. Terminology and implications: A: what do the authors mean by a "muscle-like command". What would it look like and not look like? A rubric is necessary given the centrality of the idea to the study. B: if the network dynamics represent the controlled variables, why is it considered categorically different to think about control of dynamics vs control of the variables they control? That the dynamical systems perspective better accounts for the wide array of single neuronal activity patterns is supportive of the hypothesis that dynamics are controlling the variables but not that they are unrelated. These ideas are raised in the introduction, around lines 39-43, taking on 'representational perspective' which could be more egalitarian to different levels of representational codes (populations vs single neurons), and related to conclusions mentioned later on:<br /> It is therefore interesting that the authors arrive at a conclusion line 457: 'discriminating amongst models may require examining less-dominant features that are harder to visualize and quantify'. I would be curious to hear the authors expand a bit on this point to whether looping back to 'tuning' of neural trajectories (rather than single neurons) might usher a way out of the conundrum they describe. Clearly using population activity and dynamical systems as a lens through which to understand cortical activity has been transformative, but I fail to see how the low dimensional structure rules out representational (population trajectory) codes in higher dimensions.

4. Is there a deeper observation to be made about how the dynamics constrain behavior? The authors posit that the stacked elliptical neural trajectories may confer the ability to change speed fluidly, but this is not a scenario analyzed in the behavioral data. Given that the authors do not consider multi-paced single movements it would be nice to include speculation on what would happen if a movement changes cadence mid cycle, aside from just sliding up the spiral. Do initial conditions lead to predictions from the geometry about where within cycles speed may change the most fluidly or are there any constraints on behavior implied by the neural trajectories?

5. Could the authors comment more clearly if they think that state space trajectories are representational and if so, whether the conceptual distinction between the single-neuron view of motor representation/control and the population view are diametrically opposed?

Reviewer #1 (Public Review): 

Within this manuscript the authors set out to determine the biogenesis of c-type cytochromes in methane metabolism. Compared to the bacterical cytochrome c assembly system, genes like ccmD, ccmH and ccmI are not found in archaea that contain a functional cytochrome Cs. They show that the proteins encoded within the ccmABCEF operon of Methanosarcina acetivorans are both essential and sufficient for cyt c biogenesis. They also show the substrate specific role of the mmcA cytochrome C. The authors do this using a combination of genetic, molecular, and physiological and biochemical analyses. 

The manuscript is well describes a clear set of experiments and the authors are successful in determining the biogenesis of c-type cytochromes in methane metabolism. The manuscript is well written and is easy to read.

Reviewer #2 (Public Review): 

The manuscript by Gupta et al investigates the cytochrome c maturation proteins in archaea, more specifically in methanogens. The authors demonstrate with different deletion mutant and overexpression studies that the genes ccmABCEF are required for cytochrome c maturation. By using phylogenomics, they provide evidence for several horizontal gene transfer events from bacteria to archaea giving rise to the cytochrome c maturation machinery in different clades of archaea. The manuscript is interesting and well written. The experiments are well documented and support the claims made by the authors.

Reviewer #1 (Public Review): 

The work addresses an important issue: whether subjects with cochlear implants use the same neural resources for speech listening as normal listeners. In order to address this fMRI is not possible and in this work the authors use a high-resolution form of fNIRS which measures superficial-cortex activity related to blood flow. The data implicate the use of a part of dorsolateral prefrontal cortex by CI users in the task which is not part of the specialist language network: the area has previously been suggested to be part of the multiple demand network involved in tasks including spatial working memory. The work supports additional neural resources for listening in CI users.

Reviewer #2 (Public Review): 

The authors have conducted an interesting study looking at the brain mechanisms of speech comprehension in cochlear implant users. This is a relatively under-explored area, due to difficulties in neuroimaging in this population, and so further data are welcome additions to the field. The study employs a technically challenging technique that allows the measurement of blood flow that overcomes some of these difficulties. The experimental approach is powerful, and is hypothesis-driven with respect to exploration of brain mechanisms beyond a focus on assumed recruitment of a language network, and adds a cognitive measure of working memory to create a functional ROI upon which to determine the shared recruitment of this mechanism in more difficult speech understanding that exists in the cochlear implant population. The addition of a well-matched control group allows for stronger conclusions with respect to the findings, which demonstrated that a working memory task and a word listening task shared activation in a similar region in the frontal cortex, which was greater in the cochlear implant group. This finding supports the authors' hypothesis that additional cognitive mechanisms are recruited in the cochlear implant population in order to understand speech. Research with the cochlear implant population is difficult and the authors do a good job to limit the recruitment to a relatively homogeneous population to try and decrease variance related to population characteristics. 

There are some limitations that influence the conclusions that are drawn in the paper currently. Notably, the authors hypothesize the recruitment of an anatomically defined brain region, yet the study uses a functional ROI definition. Further, this functional ROI is defined by a spatial working memory task, which is presumably used so as to be a non-linguistic task to strengthen the conclusion that general cognitive processes, like spatial working memory, are involved in difficult speech understanding in cochlear implantees. However, based on the findings the authors conclude that the brain region dorsolateral prefrontal cortex (DLPFC) is recruited in this case. Scrutiny of the DLPFC ROI shows that much of the functional activation includes the inferior frontal gyrus, which is not classically considered part of the DLPFC, inviting speculation that the spatial working memory task included cognitive mechanisms that might be assigned to the inferior frontal gyrus, for example, speech processing. Notably, the group difference in activation is circumscribed to the inferior frontal gyrus. The point here is not to debate about the localization of function to specific brain regions, rather it is to invite the authors to change their conclusions about the involvement of a brain region that seems to be minimally indicated in the results which statistically contrast the cochlear implant and control groups, and instead comment with respect to the task used. 

Another difficulty relates to the brain imaging technique employed. While there is considerable difficulty in achieving the excellent quality of data demonstrated, some methodological limitations may impact the conclusions. These relate to the extent of the field of view to cover the extent of the DLPFC, which is minimal in this case. Further, the coverage is on the edge of the field of view, where the method may have limitations in signal, and the inherent resolution of the technique and patterns of the data do not allow a strong conclusion about the exclusion of one brain region in preference for an adjacent one that is hypothesized, i.e., inferior frontal gyrus and DLPFC. Finally, the measurement from areas that overlie the cochlear implant transducer is missed, and so has the potential to influence conclusions about activation in this area. Indeed the pattern of results may indicate a finding of signal loss in the right auditory cortex. 

Findings in this study will help expand our understanding of the difficulties faced in speech understanding in the cochlear implant population, and how additional brain mechanisms may be recruited to complete this task.

Reviewer #3 (Public Review): 

Sherafati et al. investigated the brain networks used for speech perception (words in quiet) in a group of cochlear implant (CI) recipients and a control group (matched for age and gender) using high-density diffuse optical tomography (HD-DOT). Based on region-of-interest (ROI) analyses, the authors reported that the CI group showed reduced brain activity in the right auditory cortex and increased activity in the left dorsolateral prefrontal cortex (DLPFC), in comparison to the control group. 

As more common imaging methods (e.g. fMRI, EEG) are not suitable for imaging CI recipients, the use of HD-DOT is a strength of this study. The authors have been open in the interpretation of their results and acknowledged that additional data are required to support some of their arguments. The manuscript has been prepared to a high standard, in particular, the figures are clear and helpful. The authors have included a lot of supplementary materials to help readers understand and evaluate their study. 

Limitations of the study include: 

1. The rationale for some aspects of study design are missing <br /> For example, the 'critical' result of increased left DLPFC activation in the CI group was based on a ROI analysis, which was not well-motivated. 

Different brain regions support speech perception, including some domain-general brain areas. Fig. 4C suggests that CI > controls contrast identified quite a few group differences. It is unclear why the authors decided to focus on DLPFC as the domain-general brain region. Furthermore, why left DLPFC only (the authors decided on this focus by defining functional ROIs that included bilateral auditory ROIs and an ROI in left DLPFC)? 

As noted by the authors (Discussion, page 13), previous work has identified differences in activation levels between CI users and NH listeners in the right anterior temporal lobe and left middle superior temporal lobe (Zhou et al. 2018b). 

2. Results of t-tests and correlation analyses do not appear to have been corrected for multiple comparisons <br /> The authors report 3 ROI analyses and 4 correlation analyses but the associated p-values appear not to have been corrected for multiple comparisons. This issue is important because the p-value reported for the 'main' ROI result (increased DLPFC activation in CI users) is p = 0.03 (page 11). 

3. No behavioral measures were collected during the HD-DOT data acquisition <br /> The mean speech perception score (as measured by AzBio sentences) was significantly poorer in CI group relative to controls. However, this speech measure is not necessarily representative of speech perception during the HD-DOT scan (AzBio sentences vs. 15-s blocks of words in quiet). Ideally some behavioral data would have been collected during the HD-DOT scan, which could then be used to help interpret differences in brain activity

4. Results are discussed in the context of listening effort but the authors did not measure listening effort <br /> The authors explain some of their results (increased DLPFC activation in CI users) in terms of increased listening effort (e.g. Page 14, Conclusions "...and provide a potential framework for the effort that many CI users need to expend during speech perception..."). As the authors did not include any measures of listening effort, it is not clear that increased listening effort is the correct explanation for the increased activation in the left DLPFC.

Reviewer #1 (Public Review): 

Kruse and Herzschuh apply LAVESI, a machine-intensive and spatially-explicit simulation of the life-history of individual Siberian trees at the tundra-forest boundary, to call attention to the rapid reduction in the tundra biome as climate warming pushes forests toward the Arctic Ocean. The videos show the main simulation results succinctly. 

The life-history parameters of growth, reproduction, dispersal, establishment, and mortality are apparently tied to temperature, wind, and precipitation; however, the connections of life-history traits to these critical environmental variables does not appear fully described, except to state that growth is tied to temperature. 

If space is limiting in the manuscript's Methods, some of the description of machine computations could be reduced and a fuller description of how warming, wind, and water are included in the model parametrization (rather than citing a previous paper behind a pay-wall) can be provided. For instance, Figure 3 is both a computational and a conceptual graphic. Many readers might prefer to understand how climate change is incorporated into LAVESI conceptually, at least as much if not more so than how much computational time is required to run it.

Reviewer #2 (Public Review): 

This detailed modelling study provides important insight into longterm treeline advance into Siberian tundra ecosystems, quantifying the dramatic loss of tundra area of 70% even under ambitious mitigation scenario RCP2.6 by the middle of the millenium. It highlights considerable risk of extinction esp. of cold-climate tundra types. 

Strengths:

1. Emphasizes non-equilibrium of treeline position with climate conditions. 

2. demonstrates lead-lag effect of climate and treeline shift under warming, but also cooling conditions. The very slow recovery of tundra even under late millenium cooling highlights urgency of combating climate warming quickly 

3. Quantifies tundra loss, regions and speed of loss, highly relevant for science-based tundra conservation policies 

Weaknesses:

1. Systematically discussing in the introduction or the appendix main limiting factors of tree establishment and growth relevant for the study area, and mentioning those finally implemented in the model would add considerable value (i.e. limiting factors that prevent tree establishment and growth, permafrost degradation, soil nutrient development, biotic interactions (herbivory)). 

This discussion would increase traceability of methods and assessment of relevance of results, but also further emphasize how much this study is an improvement over previous studies by including some of this processes largely neglected earlier. Some of this very relevant information is mentioned in the response letter, but only partly introduced in the revised manuscript. 

2. Discussion of limitations of the modelling study is largely missing, including the following aspects: 

- Is this vegetation model coupled with a climate model? If not, feedbacks of forest expansion with climate and permafrost are currently neglected. The model is tested along gradients in selected regions, but it remains uncertain if space-for-time approach will hold in the future and further north, esp. when large-scale feedbacks are included. <br /> - What about disturbances and extreme weather conditions that might regionally impact treeline advance or tree survival? E.g. increasing tundra fire activity might strongly impact vegetation development. Also droughts/flooding might lead to regional vegetation impacts, esp. at seedling stage. Extreme events and related disturbances are predicted to increase under climate change and a discussion on how they might impact predictions by the model is needed. Are these factors all only short-term and neglectable compared to the long-term perspective modelled? If yes this should be mentioned. 

3. Figures and their legends need to be checked and improved.

Reviewer #1 (Public Review):

The authors describe a single molecule technology to identify RNA modifications. The methodology was validated with yeast ribosomes using depletion of the two major snoRNPs classes. The authors were able to resolve ribosome populations with a a single modification difference and identified which nucleotides are modified in a concerted fashion in the wt ribosome population. Based on the analysis of rRNA from the helicase mutant strains, the authors suggest a hierarchical model for the action of Dbp3 and Prp43/Pxr1, which provides an important insight in the mechanism of ribosome biogenesis. They also found that most annotated modifications do not change much upon stress or in the presence of ribosome inhibitors. These results solve several outstanding questions in understanding how potential ribosome heterogeneity and argue against the proposed involvement of rRNA modifications in stress response. The methodology can be used for other classes of RNA, which is important in view of the current interest in RNA modifications and their role in epitranscriptomic regulation.

Reviewer #2 (Public Review):

rRNA modifications have been proposed to be a main source of ribosome heterogeneity, and there has been much speculation of how co-occurrence of modification defects could both further exacerbate the heterogeneity, as well as amplify functional differences. Moreoever, there has been speculation about changes in the modification in response to different cellular states. Bailey et al directly address these questions by sequencing entire rRNA molecules using nanopore sequencing. The data not only show that most residues are modified to very high extent, but also demonstrate that most sites are independent of each other. Nevertheless, the authors do demonstrate some modification sites that are dependent on others. Some of these are readily explained by a shared snoRNA guide, but others are not. E.g., modification of the exit tunnel is concerted. Whether this is due to concerted modification, or preferential stabilization of fully modified RNA is not explored, and to this reviewer this is not necessary.

Importantly, the authors do not find any evidence for a dynamic regulation of the modifications, which to this reviewer makes sense, because rRNAs are just too long lived for this to make sense as a way to respond to cellular stress.

Overall, the claims in the manuscript are supported by data, and they are interesting and novel. I have only very minor concerns, although I am not an expert in the nanopore technology, the computational analysis, or the machine learning part.

Reviewer #3 (Public Review):

In this study the authors developed a novel strategy to map nucleoside modifications by using Nanopore sequencing of the 25S and 18S rRNAs in yeast. By comparing Nanopore sequencing reads on in vitro transcribed RNAs and RNAs extracted from cells, the authors were able to identify all 110 annotated modifications in single, full-length ribosomal RNAs.

Overall, this is an impactful manuscript that informs the field on a new technique to detect rRNA modifications and offers important insights into subpopulations of ribosomes that are lacking certain modifications. The main highlights of this paper are (1) the single molecule, direct RNA sequencing approach to detect individual modifications along an entire rRNA molecule, (2) rRNA modification is coordinated at certain positions, , and (3) subpopulations of ribosomes accumulate that are missing one or more modifications. This manuscript is relevant from the perspective of ribosome assembly, in that it informs on the order and dependencies of rRNA modifications, as well as other factors (Dbp3 and Pxr1) that are necessary for proper modification. It is also relevant in the context of the "specialized ribosome" hypothesis by showing that ribosomes are heterogenous in modification status. Most of the nucleoside modifications analyzed are promoted by guide snoRNAs, and genetic depletion of protein components of the guide snoRNPs or knockout of guide snoRNAs result in the expected decrease in the modification profiles for most positions. Interestingly they show that 2'-O-methylation is largely independent from Pseudouridylation. Another important finding of the study is the correlation of modifications at distant sites that correspond to functionally important regions of the ribosome.

I found that most of the conclusions made by the authors are supported by the data, with the exception of a few experiments described below. This manuscript will represent a major resource for the community, as it provides a new standard and approach to map ribosomal RNA modifications on single rRNA transcripts, and I anticipate that it will become a widely used tool for the scientific community. Besides the technological innovation, the information obtained on the correlation of modification at specific positions is an important finding for the fields of ribosomes and translation. In terms of specificity of identification of modifications,

The only weakness of the manuscript lies in some of the genetic experiments used to assess the impact of the inactivation of specific factors or environmental conditions on modification patterns as described below - I found three specific issues.

The first issue is the use of the prp44 cold-sensitive (cs) mutant. The authors compare the modification patterns obtained for this cs mutant after a shift to non-permissive temperature. However, there is no control experiment done with a wild-type strain shifted to the same cold temperature, which is problematic as a basic control is missing. So it would be necessary to perform a control experiment with a wild-type strain shifted to a similar temperature. Also the dbp knockout analysis is performed at steady state while prp43-cs is a cold shift so it is quite difficult to compared the result directly.

Another issue that may need to be considered is the level of depletion of individual snoRNAs after depletion of the snoRNP proteins. It is possible that some snoRNAs are depleted more rapidly than others, and that this may affect the modification patterns. The authors should perform RNA sequencing of RNA samples used after depletion of Cbf5 or Nop58 such that they can directly correlate snoRNA levels to modification levels. Unless the authors provide these data, it is difficult to conclude whether specific sites are more or less resilient to genetic depletion of snoRNP proteins.

Finally, the title of the last section of the results is also misleading in terms of its conclusions ("Resilience of rRNA modification profiles to splicing perturbations and environmental treatments"). Regarding splicing perturbations, and with the exception of the dbr1 knockout, the mutants used in the study do not result in a major depletion of intron encoded snoRNAs so it is quite expected that there is no loss of modification at these positions. Similarly, the environmental stresses used are short, and are not expected to affect modification patterns in a major way considering the stability of ribosomes. Unless the authors perform sequencing on rRNAs synthesized after a shift into stress conditions, it is misleading to state that rRNA modification profiles are unaffected by environmental treatments. My feeling is that the paper is significant enough without the studies presented in the last paragraph, and that this paragraph and the data within should be removed from the manuscript because they are inconclusive, and the title is misleading.

I spent the last few paragraphs highlighting some of the issues that need to be addressed, but overall, I found that the article presents a major advance in the field and that it provides a landmark study in our understanding of nucleoside modifications in rRNA.

Reviewer #1 (Public Review): 

The interaction or potential correspondence between chondrocranial elements and dermal skull bones has been in debate for decades. Based on 3D reconstructions of samples of laboratory mouse chondrocranial anatomy for embryonic days 13.5- 17.5, the authors reveal an embryonic relationship between the chondrocranium and forming bones of the dermatocranium in vertebrates, and provide critical data to demonstrate the role of the chondrocranium in normal craniofacial development. This will add to our understanding of the correspondence between chondrocranial elements and dermal skull bones, and the homology of dermal skull bones across major vertebrate groups.

Reviewer #2 (Public Review): 

The objective of the research performed was to assess the role of the chondrocranium in directing the morphogenesis of the skull. In this work a new method was developed to visualize the cartilages that comprise the chondrocranium, which is used to assess how these cartilage elements change over time in wild type embryos and in embryos harboring a mutation that causes premature suture closure-a model of craniosynostosis. The major strengths of this work are the combination of imaging and quantitative analyses that drive the conclusions that the chondrocranium and dermatocranium form an integrated unit of development and alterations in the chondrocranium drive changes in the dermatocranium. The methods appear robust and the authors significantly advance the field by providing results that strongly support these conclusions. Prior to this work, the changes in the skull associated with craniosynostosis were attributed to alterations in osteoblasts that form the sutures, but this work provides significant evidence that dysmorphology of the skull occurs earlier and is due to altered molecular signaling in the chondrocytes that form the chondrocranium.

Reviewer #3 (Public Review): 

The aim of the study is to tangle the over 400 million years' cooperation between chondrocranium and dermatocranium development. Mice with Crouzon syndrome were chosen for the study. The strength of this study is the novel application of machine learning techniques to segment the mouse cranium, which can be applied to a variety of vertebrates. The figures are very appealing. The major drawback of this study is that it only focuses on the Curzon mouse, even though the goal of the study is to investigate the relationship between the chondrocranium and dermatocranium. The authors emphasize that this study was undertaken to study the 400-million-year history of the cranium of Osteichthyes, which includes bony fishes, amphibians, lizards, and birds, in addition to mammals. In order to "untangle the over 400 million years' cooperation between chondrocranium and dermatocranium" as the title states, it is too obvious that they must include bony fish, amphibians, lizards, and birds. It is also unclear throughout the manuscript why the study of Curzon mice would provide insight into the relationship between the chondrocranium and dermatocranium. This study is only a descriptive study of the Curzon mouse and does not provide any insight into the "evolution" of the chondrocranium and dermatocranium. The results appear to be too much exaggerated. Again, it needs to be clearly stated why the cranial suture model is suitable for discussing the association between the chondrocranium and dermatocranium. 

There is also a need to cite and review work in the fields of evolutionary anatomy and palaeontology; it is a shame that the authors ignore important contributions by evolutionary anatomists such as Parker, Wolfgang Maier, Sánchez-Villagra, and Koyabu. In its present form, it has little relevance to evolutionary biology. 

Their conclusion that chondrocranium and dermatocranium development are associated is also not a novel finding, either. Apert mouse which exhibit the same abnormality previously reported showed that chondrocyte-specific changes in Fgfr2 alone produce an Apert-like cranial morphology suggesting that changes in Fgfr2 expression in chondrocytes may lead to the formation of membranous bone. It has already been reported that changes in Fgfr2 expression in chondrocytes have a significant effect on overall cranial morphology, including membranous bone. This study neglects such previous studies and exaggerates their results. 

This study suffers from data uncertainty because raw data was not provided. "3D coordinates of landmark data...will be made available to interested parties upon request." These raw coordinates MUST be fully provided as supplementary material, otherwise no one can re-evaluate their results. I am so surprised that even basic statics (PC scores, loadings, eigen values, explained variance) are not provided. Data availability and transparency are very important. 3D models are also not provided in the review, so at this point I cannot be sure of the accuracy of their segmentation. They have stated that they will make it available at https://www.facebase.org/ and/or https://scholarsphere.psu.edu/, but it should be accessible now for reviewers. Facebase is fine, but it should not be provided on their own institute's server that may go out of service at any time. It should be provided through a permanent public archive.

Reviewer #1 (Public Review):

This study aimed to identify the genetic foundation favoring common, nearly predictable selection of lasR mutants in laboratory and clinical isolates from persons with CF. They selected these mutants using a predictable and quantitative framework of evolution experiments and then identified their genetic underpinnings by a a suppressor screen. The role of cbrAB as a key intermediate is important and ties together several reports of nutrient-dependent advantages of lasR like phenylalanine, including those reported recently (Scribner et al JBact 2021).

The metabolomic study is interesting and offers a plausible correlation between the evolution of lasR mutants during infections of pwCF and the nutritional conditions that select these mutants. Naturally, these are not causative, which should be clarified. The summative figure describing a model of metabolic and hence genetic diversity of PA is also elegant.

The figures and writing are clear and of high quality.

Reviewer #2 (Public Review):

In this paper, the authors thoroughly explore the selective advantage of LasR- mutants of Pseudomonas aeruginosa. As the authors state, selection of loss of function mutations in quorum sensing regulators, including LasR, is frequently observed during chronic infections and laboratory culture, but the drivers of this selection are poorly understood. Mould et al. utilize mathematical modeling, evolution experiments, and whole genome sequencing to show that metabolic advantages are sufficient for selection of LasR- mutants. Further, the authors use a reverse genetic screen paired with evolution experiments to identify the CbrA/CbrB pathway as necessary for this selection. Subsequently, the authors characterize the roles of genes within this pathway with regard to LasR- phenotypes. The authors also determined the nutrients enriched in bronchoalveolar lavage fluid from people with cystic fibrosis and show that LasR- strains have advantages in this nutrient environment. The authors' conclusions are well supported by their data and thoroughly verified using complementary approaches. In addition, the authors provide extensive supplementary data exploring alternative hypotheses related to their findings.

There are several notable strengths of this work. For instance, the authors performed many experiments using both the PA14 laboratory strain and a cystic fibrosis isolate to illustrate the applicability of their findings to distinct genetic backgrounds. In addition, the authors' use of a mathematical model to test the hypothesis that metabolic advantages of LasR- mutants are sufficient to explain their selection and their application of a reverse genetic screen to evolution experiments are particularly clever approaches.

The authors' finding that lasR mutations arise less frequently on a ∆cbrA or ∆cbrB mutant background is very interesting. Also, among the most compelling findings of this study was the parallel evolution of mutations in the downstream crc gene in ∆cbrB mutant cultures. Together, these findings strongly suggest that increased CbrB expression of lasR mutants plays an important role in their selection, as stated in the paper.

Reviewer #3 (Public Review):

The work of Mould et al. focuses on a protein LasR, which is a transcription factor involved in quorum sensing in Pseudomonas aeruginosa, which can frequently cause disease in patients with Cystic Fibrosis (CF). Isolates with loss-of-function mutations are frequently found in both environmental and clinical samples, and are associated with more severe outcomes during infection in people with CF. The authors set out to determine why strains with these loss-of-function have a seeming advantage over wild-type (WT) cells, both based on growth and mechanistically. They use mathematical modeling, experimental evolution, sequencing, and metabolome analysis to come to their well-supported conclusions. They determine that LasR- mutants can quickly take over cell populations when competing with WT cells using serial passage. Using reverse genetics, they then identify a pathway which contributes to this advantage. They ultimately determine that LasR- mutants alter metabolism in a way that they can grow on compounds most commonly found in the lungs of patients with CF via the CbrAB pathway.

The conclusions in this paper well-supported, and the experiments mainly add to these conclusions. These methodologies and conclusions will add to the evolution field by helping to understand more about why certain genetic changes give an advantage to cells, even when there may also be disadvantages associated with those mutations. However, there are a few passages in the writing which confuse the conclusions a bit, and there a few places in the writing where it is unclear that the comparisons between cultures are done using the same methods. Specifically:

1. It is not clear whether or why ∆anr or ∆rhlR strains are used to compare rates of LasR- mutations.<br /> 2. The logic describing why the authors expect higher activity of the CbrA-CbrB-crcZ pathway in LasR- strains, and therefore more loss of function alleles in Crc or Hfq, and then confirm this theory with the data showing that they have mutations in crc or hfq in ∆cbrA and ∆cbrB mutants (but not in WT strains), where there should not be LasR- mutations, is not clear.<br /> 3. It is not clear that all growth curves, which are compared to the mathematical model throughout the paper, are performed the same way (i.e. passaged every 48 hours).

Reviewer #1 (Public Review):

The authors use both in vitro signaling assays, knockdown in chick neural tube patterning assays and some limited use of Plexin mutant mice. The in vitro work convincingly demonstrates that misexpression of several Plexins is sufficient to enhance HH signaling in a way that depends on the Plexin GAP domain. Not addressed is how the GAP activity promotes HH signaling. The in vivo data are extremely interesting. However, alternative interpretations of the data are not assessed and need to be before the conclusions favored by the authors can be asserted.

Reviewer #2 (Public Review):

This is interesting work that expands our knowledge of Hedgehog signaling. The work is well-done, well-written, and the figures are clear. I have comments that would help strengthen some of the experiments and improve the manuscript. In particular, the in vivo loss of function experiments could be measured in additional ways (using additional endpoints) to provide a convincing case of the role that Plexins play in Hh signaling in vivo.

Specific comments, in no particular order:

1- The authors show that the effect of SmoM2 or Gli1 overexpression on Hh pathway activity can be potentiated by Plexins. They then conclude that "These data suggest that PLXNs function downstream of HH ligand at the level of GLI regulation...". It is unclear to me how this experiment allows them to conclude this, as the effect of Plexins could be downstream of Gli1, through the regulation of the transcription machinery, for example.

2- Are primary cilia formed normally and present at normal frequency in cells with loss or over-expression of Plexins? This could help understand better how Plexins act to modulate the Hh pathway.

3- Are Gli1 protein levels affected by Plexins?

4- In order to provide a convincing case for the role that Plexins play in Hh signaling in vivo, the in vivo Plexin loss of function experiments should be assessed in additional ways to Gli1-lacZ (Figure 6). Also, proliferation should be measured (as previously shown to be Hh-dependent).

5- Data showing whether Plexins bind Shh (or not) should be presented.

6- The authors show that increased Plexin activity in chick neural tubes increases cell migration into the neural tube lumen. Is this effect of Plexins Gli-dependent?

7- In the chick neural tube experiments, how can the authors conclude that Plexin promotes Gli-dependent cellular responses since their data show that Plexin is not significantly affecting the fate (NKX6.1 and PAX7) of the cells? I was confused by this. The image shows a change, but the quantification does not.

8- Could loss of function experiments in chick neural tube using RNAi against multiple Plexins be performed? This would provide a very convincing case of the requirement of Plexins for Shh signaling.

Minor points:

9- Figure 1 panels H-I need a negative control for siRNAs.

10- Figure 3B needs to control for Plxn1ΔECD expression levels (by western). Can higher activation of the pathway be explained by higher Plexin protein expression?

Reviewer #3 (Public Review):

The main strengths of this study are the compelling data derived from the use of well-established cell-based assays of Hedgehog signalling and novelty of the finding that Plexins can modulate the response of cells to Hedgehog. The experiments are well designed and carefully controlled.

The main weaknesses are as follows:<br /> 1. Plxna2 is expressed at levels lower than a3, b2 and d1, but it is not explained why this gene was knocked out in cell lines in preference to the other three.<br /> 2. Most of the analysis and the main conclusions of the study are based on the 3T3 experiments. The data supporting the in vivo significance of these findings are less strong:<br /> First, using electroporation of the chick neural tube, they revealed that constitutive Plexin activity can replicate only a subset of the effects of Gli over-expression. It would be relevant to know if ectopic cell migration can be caused by levels of Gli activity lower than those sufficient to induce Nkx6.1 expression - I am not sure if this is already known.<br /> Second, the authors investigate the consequences of loss of plexin function in the hippocampus, using mouse Plxna1 and Plxna2 mutants. This is a bit puzzling given that their own cell-based assays show that loss of either or both of these proteins has no impact on the response of 3T3 cells to ligand. Moreover, a previous study cited by the authors (Cheng et al 2001) reported that Plxna3 shows the highest and most widespread expression in the CNS and in the hippocampus in particular. Plxna2, by contrast is expressed at much lower levels whilst Plxna1 was detected principally in mature pyramidal cells. It is not clear why the authors chose to focus on these particular Plexins and to what extent the requirements for Plexin function have been rigorously tested.

Reviewer #1 (Public Review):

The authors have used a multipronged approach to elucidate the role of Growth Differentiation Factor- 15 ( GDF-15) in cardiometabolic disease.

Strengths of methods and results:

1) The use of well-defined cohorts FINRISK and INTERVAL with high-quality data

2) The use of different methods to elucidate causation - (1) observational analysis for associations of GDF15 levels with disease outcomes and quantitative biomarkers, (2) GWAS in independent cohorts and (3) mendelian randomisation, reverse mendelian randomization and effect of protein truncating variants on cardiometabolic traits.

Weaknesses of methods and results:

1) There are limitations in the plasma GDF15 assays with heterogeneity across different methods at least partly due to epitope binding artefacts due to a common missense variant. The authors performed additional analyses to deal with this.

2) The truncated variants had no significant effects but given the subjects also had a normal allele this presumably compensated for the truncated allele. A homozygote or compound heterozygote for a non functioning allele could be highly informative but will probably be rare and hard to find. This could be further discussed.

The authors largely achieved their aims and the results support their conclusions.<br /> The work is very thorough and comprehensive and is very clearly presented so that after each heading its very clear what has been found and the authors then elaborate on this.<br /> They show that plasma GDF15 levels are a non specific general marker for risk of multiple cardiometabolic and other diseases, probably an indicator of metabolic stress, analogous to CRP, and not a causative factor.

Likely impact of the work on the field, and the utility of the methods and data to the community. The well conducted and multi pronged approach is an excellent example of how research into questions of this type can be addressed. The findings provide a good basis for further research into GDF 15.

Additional significance of the work: This highlights again the critical role of the assay determining the analyte under investigation. If the assay has significant limitations the whole study has inherent limitations. The authors note this and address this as best they can.

Reviewer #2 (Public Review):

The authors attempt to explore the role of growth differentiation factor 15 (GDF-15) in cardio metabolic traits using different designs (i.e. observational analyses, genetic analyses, and Mendelian randomisation). Although the authors found that GDF-15 associated with multiple traits in a cohort design, the authors did not find evidence in favour of causation using Mendelian randomisation. Whilst the use of genetics to verify observational results should be commended and the conclusion is largely supported by the analyses presented, there remains some limitations in this study.

Strengths:

The use of both observational analyses and genetic analyses is a powerful approach to ascertain the causal role of GDF-15 in cardio-metabolic traits.

The use of different cohorts to obtain genetic predictors of GDF-15, and hence improve the statistical power of the Mendelian randomisation design. Based on genetic analyses, intervening on GDF-15 would not impact on cardio-metabolic traits although there was evidence of reverse causation for some traits such as body mass index.

Limitations:

Whilst the use of Mendelian randomisation design may circumvent issues with confounding in observational studies, there was a lack of certain analyses which may improve robustness of findings, as well as concerns over the instruments which appeared to be correlated.

Nevertheless, the use of different approaches presented in this paper would be a good example of using triangulation of evidence to assess the causal role of potential markers in cardio metabolic traits, which has implications for targets of interventions.

Reviewer #3 (Public Review):

GDF15 has been shown to be upregulated in response to cellular stress and cancer. Elevation of Gdf15 is observed in animal models of obesity and genetic overexpression of Gdf15 showed beneficial effects against obesity. Higher GDF15 levels have also been associated with all-cause mortality in humans. This study conducted a systematic and extensive phenotypic and genotypic analysis of GDF15 with cardiometabolic traits and diseases, and to ascertain the causal relationship between GDF15 levels and cardiometabolic traits using Mendelian randomization and protein-truncating variant analysis.

The strength of this study is the utilization of three large biobanks to dissect the causality of upregulated GDF15 and its association with cardiometabolic diseases, the relationship of GDF15 levels with commonly analyzed blood biomarkers, and GWAS variants with GDF15 levels. The other strength is that the authors performed rigor analysis by taking into consideration the different methods of measuring GDF15 levels in different cohorts and utilizing protein truncation variants as a valid approach to determine the function of GDF15. The presentation of their findings is very good, for example, the forest plots of Cox proportional hazard for independent predictors of targeted outcomes gave readers a clear picture for the GDF15 contributions.

The authors' conclusion is justified based on the comprehensive analysis performed. This is the first study that utilizes the largest sample size of data and the comprehensive genetic, GDF15 levels and potential causality evaluation.

Reviewer #1 (Public Review):

This study focuses on elucidating the function of CD59, a small GPI-anchored glycoprotein, in Schwann cell development. Patients with CD59 deficiency suffer from neurological dysfunctions, but the link between CD59 deficiency and the development of neurological dysfunctions remains unclear. To clarify this link, the authors used zebrafish as an animal model. They generated cd59 mutant zebrafish and studied their Schwann cell development. The authors started this study by showing CD59 expression data from different sources in the Schwann cell and oligodendrocyte lineages in zebrafish and mice. They continued by demonstrating that CD59 is expressed only by a subset of developing Schwann cells, which is very interesting conceptually for the identification of different Schwann cell populations and their specific functions and also for the potential development of future techniques targeting specific Schwann cell populations. However, since the authors focused in the following parts of the article on Schwann cell development, it is unclear why they have included data on oligodendrocytes at the start of the manuscript.

In this study, the authors show that cd59 ablation in zebrafish leads to increased Schwann cell proliferation between 48 and 55 hpf (hours post fertilization), which is quite convincing. However, they claim that this transient increase in proliferation leads to impaired myelination and node of Ranvier formation. Unfortunately, these findings remain correlative and it appears unclear why an increased number of Schwann cells that stop proliferating at the same time-point as wild type Schwann cells would impair myelination and node of Ranvier formation. This phenotype is attributed by the authors to increased proliferation of Schwann cells between 48 and 55 hpf, which seems rather unlikely or not supported by the data currently presented. The hypomyelination phenotype is rather mild, while the impairment of node of Ranvier formation seems quite strong - however, the data currently presented is not very convincing and needs improvement. The data showing an increase of complement activation in cd59 mutants is also not very convincing and should be improved. In addition, the link between increased complement activation and increased proliferation remains to be proven in the context of this study, and the choice of dexamethasone as an inhibitor of complement activation does not appear to be the best choice since it is not specific to the complement.

Page 49, lines 437-439: Here the authors claim that their data "demonstrates that developmental inflammation aids in normal SC proliferation and that this process is amplified when cd59 is mutated." The data presented in Figure 6C-D and commented by the authors on page 49, lines 435-437, show however that "Dex treatment in cd59uva48 mutant embryos restored SC numbers to wildtype levels, whereas wildtype SCs were not significantly affected by Dex application". Dex (dexamethasone) was used here to inhibit inflammation and associated complement activation. Therefore, these data do not show that developmental inflammation aids in normal SC proliferation, but rather that it has no influence.

Dexamethasone treatment: The authors claim that dexamethasone treatment, by decreasing inflammation and associated complement activation, leads to a decrease of SC proliferation in the cd59 mutant. To support this, there is only Figure 7-Figure Supplement 1 showing a decreased SC number in the mutant treated by dexamethasone as compared to vehicle-treated mutant. To strengthen this point, the authors also need to specifically quantify proliferation by EdU incorporation, as they did in Figure 4, and also cell death.

In addition, the mechanistic hypothesis of increased proliferation in cd59 mutant is that cd59 interferes with the activation of the complement and complement-induced pore formation in the plasma membrane. However, dexamethasone is not a specific inhibitor of the complement. Therefore, its potential effect on SC proliferation could be due to other effects than complement inactivation. It is unclear why the authors did not use an inhibitor of the complement that is more specific than dexamethasone.

Page 54, lines 456-457: The following statement "Collectively, these data demonstrate that inflammation-induced SC proliferation contributes to perturbed myelin and node of Ranvier development." is not accurate since these data remain correlative. Indeed, there is in this study nothing showing that increased SC proliferation between 48 and 55 hpf leads to perturbed myelin and node of Ranvier development. In addition, the term "inflammation" is not precise enough here. What the authors attempt to show is an increase of complement activation due to the absence of cd59 expression in SCs. The authors did not try to induce inflammation in wild type animals to see whether this induces proliferation and perturbed myelin and node of Ranvier development. They also did not try to directly knock down C8/C9 in cd59 mutants to see whether they would rescue the phenotype of the cd59 mutant, at least to some extent. In addition, their statement mentioned above needs to be more precise by stating that their findings apply to cd59 mutants and not to wild type animals.

Reviewer #2 (Public Review):

Wiltbank et al. investigate the functional relevance of cd59 expression, mainly focusing on the developing peripheral nervous system of zebrafish. By analyzing pre-existing bulk and scRNA-seq datasets they find that cd59 is expressed in both myelinating Schwann cells and oligodendrocytes, the myelinating cells of the peripheral and central nervous system, respectively. Indeed, CD59, a small GPI-anchored glycoprotein, is an abundant myelin protein in the CNS of adult zebrafish according to prior proteome analysis. The authors first generate a zebrafish reporter line to further characterize cd59 expression and then by genome editing (CRISPR/Cas9) generate a loss of function line. By CISH and FISH they validate the RNA-seq data and observe expression of cd59 in a subset of developing myelinating Schwann cells (SCs) in the LLN. They find that CD59 restricts the over-proliferation of Schwann cells and affects the structure of myelin and the nodes of Ranvier. Considering prior knowledge that CD59 plays a role in suppressing the complement system, the authors assessed how the complement reacts in cd59 mutant zebrafish. Strikingly, they found that Schwann cells were not protected from complement attack when CD59 was lacking. Inhibiting inflammation using Dexamethasone not only restored Schwann cell numbers to wildtype levels but also improved Node of Ranvier clustering and myelin volume. Hence, the authors conclude that cd59 expression protects developing Schwann cells via inhibiting developmental inflammation.

Generally, the conclusions of the manuscript are well supported by the data. A few minor points regarding phrasing and figure design could be further improved, but overall this is an interesting study that could also become relevant for future therapeutic translation. Together, the study presents a relevant and novel topic and will find an interested readership in the communities working on myelinating cells, the complement system / innate immune system, and their interactions.

Reviewer #3 (Public Review):

Wiltbank and colleagues explore the function of CD59 in developmental Schwann cell myelination. Using previously published transcriptomics data sets they arrive at CD59 as a differentially expressed gene in myelinating glia. In addition, patients with pathogenic variants have neuropathy. The authors construct a transgenic zebrafish reporter line for cd59. Surprisingly, it labels a very, very small percentage of Schwann cells (less than 10% throughout development). The authors then construct several loss-of-function mutants for cd59. They report these mutants have increased numbers of Schwann cells, but nerves are smaller and EM shows they have reduced the number of myelin wraps. Consistent with impaired myelination they also observe fewer nodes of Ranvier. The authors suggest loss of cd59 results in increased MAC deposition on myelinating Schwann cells. Remarkably, using an inhibitor of inflammation (dexamethasone), the authors show that they can normalize/rescue the main phenotypes: 1) normalize the number of SCs, 2, dramatically improve myelination to normal nerve volumes, and 3) rescue node of Ranvier formation. This last experiment that rescues the phenotype is really terrific. The experiments are mostly very well done and the story is both interesting and conceptually novel. Nevertheless, there are a few points that I think the authors could address:

1. It is very surprising that the cd59 reporter line only showed expression in a small subset (10% or less) of Schwann cells. How do the authors explain the widespread effects? Similarly, the authors make a point of stating that motor Schwann cells did not express cd59. Did myelinated motor axons show the same phenotype - reduced myelination, impaired node formation? How can the expression of cd59 in only 10% of cells cause widespread effects throughout the nerve? How can it limit overproliferation if 9/10 cells don't even express it?<br /> 2. It is surprising to me that there is a significant increase in SC proliferation, but no change in the length of myelin sheaths. Does this mean there are more SCs that remain unmyelinated and undifferentiated?<br /> 3. The results showing deposition of the MAC (via C5b-9+C5b-8 immunostaining) are not convincing. The overall background level of immunostaining is dramatically increased. This result is central to the overall story in the paper. What controls were performed to confirm this doesn't simply reflect an overall higher background artifact during immunostaining?<br /> 4. Can the authors speculate on a mechanism for how promoting more MAC results in increased proliferation?

Reviewer #1 (Public Review):

Liu et al investigated the role of Wnt/β-catenin pathway in the genesis of thermogenic adipocytes. Their study shows that some adipocytes exhibited Wnt/β-catenin signaling ("Wnt+ adipocytes") in intrascapular brown adipose tissue (iBAT), inguinal white adipose tissue (iWAT), epidydimal WAT (eWAT), and bone marrow (BM). There was a different level of the possession of Wnt+ adipocytes between the different depots with iBAT expressing 17%, iWAT expressing 6.9%, and eWAT expressing the least at 1.3%. Expression of these adipocytes was noted on embryonic day 17.5 and was present in a higher percentage in female mice compared to male mice and in younger mice compared to older mice, which aligns with their observation that Wnt+ adipocytes are thermogenic.

The authors also noted that Wnt+ adipocytes can differentiate from human stromal cells. In regards to the pathway, Wnt/β-catenin adipocytes are distinct from classical brown adipocytes at molecular and genomic levels. It was noted that Tcf7L2 was largely expressed in Wnt+ adipocytes but other Tcf proteins (Tcf 1, Tcf 3, and Lef1) were not. Wnt- cells showed a reversible delay in maturation with LF3, however, no cell death was noted. Wnt/β-catenin adipocytes seem to depend on AKT/mTOR signaling. It was further shown that insulin is a key factor in mTOR signaling and Wnt+ adipocyte differentiation.

Upon cold exposure, UCP1+/Wnt- beige fat emerges largely surrounding Wnt+ adipocytes, implicating that Wnt+ adipocytes serve as a "beiging initiator" in a paracrine manner. Lastly, mice with implanted Wnt+ adipocytes had a significantly better glucose tolerance which suggests that Wnt+ adipocytes have a beneficial impact on whole-body metabolism. I found no major flaws in the method and data largely supports their conclusion that Wnt+ adipocytes have (at least some) a significant role in thermogenesis/metabolism, which I think is a very impressive and innovative finding.

Reviewer #2 (Public Review):

Liu et al present evidence for the surprising finding of Tcf/Lef-active, "Wnt+" mature adipocytes. They report that Wnt+ adipocytes arise during embryogenesis and regulate cold-induced beiging in surrounding adipocytes. Tcf/Lef transcriptional activity in these cells is Wnt-ligand independent and instead appears to be stimulated by insulin-dependent AKT/mTOR signaling. Using a diphtheria toxin inducible depletion mouse model, the authors show that Wnt+ cells play an important role in glucose homeostasis.

As the authors have acknowledged, proper assignment of adipocyte nuclei is a notoriously difficult histological challenge. Mesenchymal cells sit directly adjacent to the adipocyte plasma membrane and their nuclei are often incorrectly assigned to the adipocyte both in vivo and in vitro. Pparg nuclear co-staining is helpful, however, Pparg is very highly expressed by endothelial cells and Col15a1+ committed preadipocytes, which are intercalated throughout the adipose. The authors have made an impressive attempt to address this concern by generating a Tcf/Lef-CreER mouse line to fluorescently label Wnt+ adipocytes, however, it is not entirely clear if the images presented support the conclusion that mature adipocytes are being labeled. Given that Wnt+ mature adipocytes are the core conclusion of this manuscript, and because this hypothesis runs counter to a large body of literature concluding that Wnt signaling inhibits adipogenesis, the authors have assumed a very high burden of proof that these are indeed Wnt+ mature adipocytes in vivo.

To address these concerns, the authors could utilize the specificity of in vivo single-nuclei RNA-Seq. Several data resources have been published (https://singlecell.broadinstitute.org/single_cell/study/SCP1376/a-single-cell-atlas-of-human-and-mouse-white-adipose-tissue), and the authors should re-analyze these data for subpopulations of mature adipocytes that express a transcriptional signature of active Tcf/Lef signaling. It is unfortunate that the authors were unable to successfully perform single-nuclei analysis of the Wnt+ adipocytes as this would significantly enhance this manuscript. The physiologic relevance of the single-cell analysis of immortalized, in-vitro differentiated clonal cell lines is questionable.

Reviewer #3 (Public Review):

It is becoming increasingly clear that adipocytes are not homogenous, but rather comprise several distinct subtypes with specific physiological functions. The mechanisms that underlie the development and distinct roles of each adipocyte subtype are of great interest for understanding the biology of metabolic regulation and its impairments in metabolic disease. In this manuscript, the authors describe a previously unknown population of adipocytes in mice, which are characterized by a special form of beta-catenin signaling. They perform a comprehensive series of experiments in cultured cells, in mouse models of in-vivo lineage tracing, and transplantation experiments to define the origin and function of these adipocytes. They find that the formation of these Wnt+ adipocytes is dependent on insulin signaling, and find possible roles in thermogenic adipose tissue development. Overall, the conclusions of this study are very convincing in their identification of a subpopulation of adipocytes displaying non-canonical Wnt signaling. The proposed role of these adipocytes as regulators of thermogenesis is more ambiguous, and their physiological function remains unclear.

• The new adipocyte types are identified through expression of a reporter for TCF/Lef signaling. This reporter is classically activated by Wnt/beta-catenin and using both siRNA depletion of beta-catenin as well as an allele lacking its transcriptional activation domain, the authors confirm the reporter expression is dependent on the presence of beta-catenin and TCF7L2, but independent of canonical Wnt signaling.<br /> • The involvement of TCF7L2 is also probed using a specific inhibitor of the beta-catenin/TCF7L2 interactions, LF3, which inhibited reporter expression. Inhibition of canonical Wnt signaling was without effect.<br /> • The authors isolate clonal lines of precursor cells that give rise to Wnt+ or Wnt- adipocytes from mouse brown adipose tissue. They find that Wnt+ adipocytes are dependent on the Wnt pathway, as inhibition by LF3 induces cell death.<br /> • To further probe the nature of Wnt+ and Wnt- adipocytes, the authors perform scRNASeq on cells after 7 days of adipose induction and find 2 distinctive cell populations. The finding of 2 distinct populations is expected, given the a priori separation of cells as a function of GFP expression. It is not clear why scRNASeq was chosen over RNASeq on the population, since the fat content of adipocytes may preclude full characterization of the most differentiated cells. Overall, this experiment is less informative on the mechanisms by which Wnt+ adipocytes display Wnt signaling dependency for viability, and what their functional role might be.<br /> • The non-canonical nature of Wnt signaling in Wnt+ adipocytes prompted the authors to explore the role of the insulin/PI3K/AKT/MTOR pathway. They find enhanced basal activity of this pathway in Wnt+ adipocytes. It was not explored whether this enhanced activity persists under insulin stimulation; this is relevant as feedback mechanisms within the signaling pathway may result in lower signaling under stimulated conditions.<br /> • To test the relevance of insulin signaling in-vivo on non-canonical Wnt signaling in adipocytes the authors use the Akita mouse, which lacks the insulin-2 gene and find a marked decrease in reporter activity, confirming the requirement for insulin signaling for expression of this non-canonical Wnt pathway.<br /> • To determine the functional role of Wnt+ adipocytes, the authors explore their relationship to mitochondrial respiratory activity and thermogenesis. They perform experiments to monitor mitochondrial membrane potential and oxygen consumption rate and find higher overall O2 consumption, and lower membrane potential in adipocyte populations vicinal to Wnt+ adipocytes. Overall these results are not fully convincing: The traces are highly variable from cell to cell, and rigorous quantification of uncoupled respiration is limited by the small number of cell lines analyzed; only one cell line of Wnt- and two Wnt+ adipocytes are analyzed. In situ differences in membrane potential would be more convincing if performed on homogenous collections of Wnt- and Wnt+ adipocytes to better understand stochastic variance.<br /> • To determine the role of Wnt+ adipocytes in-vivo thermogenesis, the authors expose mice to cold temperature and monitor the proportion of UCP1+ adipocytes in relation to Wnt signaling. They find a proportion of Wnt+ adipocytes expressing UCP1. Whether this proportion is higher or lower than that of Wnt- adipocytes is not quantified, so it is unclear whether Wnt+ adipocytes preferentially develop beige characteristics. The authors find that UCP1+, Wnt- adipocytes are topologically close to Wnt+ adipocytes, and hypothesize a paracrine signaling role. However, this correlation may be explained by known topological biases in inguinal fat pad beiging, where adipocytes closer to lymph node preferentially induce UCP1. The Wnt+ adipocyte population may coincidentally be present in this region.<br /> • To functionally determine the role of Wnt+ adipocytes in thermogenesis, the authors ablate the Wnt+ lineage through expression of diphtheria toxin using a Fabp4-Flox-DTA mouse crossed to Tcf/Lef-CreERT2 mice. Less than 50% of these mice displayed impaired thermogenesis upon cold exposure. The authors interpret this finding to signify a partial role for Wnt+ adipocyte beiging in thermogenic regulation. This conclusion is not fully supported, as Fabp4 is expressed in many cells other than adipocytes, and therefore the phenotype of the affected mice is not unambiguously attributable to loss of Wnt+ adipocytes. An additional concern is that diphtheria toxin-induced cell death will lead to tissue inflammation, with potential functional effects on thermogenesis. The degree of cell death and inflammation should be measured and reported.<br /> • The finding that Akita mice lack Wnt+ adipocytes was used to determine whether these mice are susceptible to cold-induced challenges. The authors report a decrease in cold-induced UCP1 expression in these mice. This conclusion, derived from a single immunofluorescence image, is not fully convincing in the absence of additional metrics.<br /> • To further explore the role of Wnt+ adipocytes in systemic metabolism, the authors conduct implantation studies of Wnt+ adipocytes and measure effects on glucose tolerance. They show a significant difference in glucose excursions in mice harboring fat pads developed from Wnt+ adipocytes. These results are convincing, but the conclusion may be due to enhanced volume of additional functional fat developing from Wnt+ adipocytes.

Reviewer #1 (Public Review):

This paper extends a previous analytical method that the authors developed to evaluate the time to infectiousness of COVID-19, in order to evaluate differences in the generation interval across different time periods during the course of the pandemic in England in 2020. The time to infectiousness (i.e. how long is it until infected individuals start producing virus in a way that is a risk of infecting others) is a generalisable concept. That is unless we expect there to be inherent differences in the way infected individuals progress to becoming infectious (when looking at distributions of outcomes, comparing between populations of interest) we can take a result from one population of individuals, and assume that it gives us a reasonable idea of how long it takes to become infectious, in another population. Differences in the way people come into contact with each other will have some influence on this, but generally speaking if a person is infectious after 4 days in China, you should be consider a person to be a risk of infecting others after 4 days in other countries as well.

In contrast, generation time (how long does it take an infected person, on average, to infect the persons they are going to infect?) depends strongly not just on the inherent characteristics of the virus, and progression of disease in individuals, but also (more strongly that time to infectiousness) the circumstances of contact between individuals. Because generation time is tied to so many other factors, one of the most reliable ways to estimate generation times is to analyse data where there are groups of in-contact individuals where there is likely to be highly likely that there is only one generation of transmission involved (where contacts between individuals are clustered, possibly two but with three generations highly unlikely). In this case, the most important unknowns are the time from when individuals are infected to when become infectious and the time to when they test positive - the requirement for time to infectiousness is why the methods used in the initial paper are appropriate for generating better generation time estimates.

As most published results relate to the very early stages of the pandemic in China where extensive contact tracing were done, there is some interest in understanding whether the generation times differ substantially in other locations and if they change over time (and therefore, why). In this analysis, Hart et al. estimate generation times across three, three month time periods using household contact data in England in 2020, and show differences in generation time estimates depending on the method used (in particular, when considering an approach which ties infectiousness to symptomatic development which they showed provided better results compared to other methods in their previous paper) and the period of 2020 over which the estimates are taken. While the result appears technically robust for the data analysed, its usefulness is limited by difficulty in extending the results - while a different dataset from ones used for the analyses in China they refer to, and from the result of Challen et al. that looked at contacts of international travellers in the UK, it is also in its own way quite specific and further breakdown of possible factors would be worthwhile. First, the limitations to household contacts means that it is not representative of general transmission in the population - household contacts are high risk, with many opportunities for transmission and may therefore be relatively short. Generalised contacts outside of households are likely to be less frequent and often of shorter duration and more strongly affected by diurnal and weekly rhythms. Second, it is also known that demographic factors such as ethnicity and income are strongly linked to infection and severe infection risk. While this does not tell us directly about any links to infectiousness and infectious contact, it is reasonable to consider a connection - and therefore a link to generation times. As such, in this relatively small sample (172 households, with much higher numbers in the first 3 months, compared to the middle or last three) differences in demographics may influence generation times as well. Finally, the alpha variant, first identified in Kent, was probably circulating for much of the final three months of this analysis - dominant by early 2021 in the UK, it would have had a variable proportion across much of those final three months, and also varied geographically in terms of proportion as well, with a much earlier rise in the SE and in London). Unless those proportions are known, it would be difficult to know how much differences in generation times are due to the variant, to demographics, or other, possibly behavioural factors. Thus some caution should be applied before taking general lessons from it, at least in the absence of those additional considerations.

Reviewer #2 (Public Review):

In this work, Hart et al infer the generation interval for SARS-CoV-2 using infector-infectee pairs from household data. The generation interval is obtained across three different time intervals (March-April, May-August and September-November) and using both an "independent transmission" model and the "mechanistic" model that was originally proposed in Hart et al 2021. The main result is that the inferred generation interval in September-November has decreased compared to the earlier months of the pandemic, irrespective of the model considered. Overall, the conclusions drawn in the paper are well supported and have been shown to be robust through a thorough sensitivity analysis.

Strengths

- They use a mechanistic model to account for the change in infectivity at symptom onset.<br /> - A major strength of this investigation is that they can observe the dynamics of the generation time over three different time periods of the pandemic. To my knowledge, this is a novel result that allows for a more up to date understanding of SARS-CoV-2 transmission.<br /> - Whilst not highlighted in the text, it appears that there has been significant effort to extend the likelihood function to appropriately model household dynamics. This is non-trivial work in my opinion, and I believe the details of the derivation will be of use to mathematical modellers that deal with susceptible depletion in their data.

Weaknesses

- The main weakness of the paper in its current form is that the analysis appears superficial, with a large amount of curve fitting and very little explanation. It would be beneficial if the authors delved more deeply into their results, especially with the mechanistic model. It would be very interesting to relate the changes in generation time to mechanisms of transmission.<br /> - The authors calculate the mean and standard deviation of the generation interval across three different time points; however, they only present one figure with the distribution of the generation time (Figure 2). It would be interesting to know how the generation time distribution changes in time, as opposed to just the mean and standard deviation. I believe that such an analysis would link nicely to their previous work, where they highlight the importance of ongoing public health measures such as contact tracing.

Reviewer #3 (Public Review):

The authors have previously published a mechanistic model for inferring infectiousness profile that explicitly models dependence of the risk of onward transmission on the onset of symptoms on an individual. In the present study, they apply this model as well as another more commonly used model which assumes these two things (transmission risk and onset of symptoms) to be independent, to data from a household study conducted from March-Nov 2020 in the UK. Both the models find that the mean generation time in Sept-Nov 2020 is shorter than in the earlier periods of the study.

This is well-presented study with careful analysis and extensive sensitive analysis which shows that the modelled estimates are robust to a range of assumptions.

Reviewer #1 (Public Review):

The Non-structural protein (Nsp)-1 from SARS-CoV2 mimics the binding mode of eukaryotic initiation factor 3 (eIF3j) to the mRNA entry tunnel of the 40S ribosomal subunits and blocks the entry of mRNA, which shuts down host protein synthesis. As a result, the host immune function is suppressed. This makes Nsp-1 an attractive target. The manuscript reports on the repurposing of Montelukast, an FDA-approved drug, which forms a complex with the C-terminal helices of Nsp-1 and prevents it from binding to the mRNA channel on the 40S ribosomal subunit. The drug displays a binding affinity (KD) of 10.8{plus minus}0.2 μM in vitro and reverses the effects of Nsp-1 on host protein synthesis. The authors demonstrate this using the firefly luciferase reporter gene assay and by observing a reduced expression of viral spike protein in HEK-ACE2 and Vero-E6 cells. Short molecular dynamics simulations are also used to study the binding mode of Montelukast to Nsp-1.

Reviewer #2 (Public Review):

In this manuscript Afsar et al. report on a drug repurposing effort to target the non-structural protein 1 (Nsp1) of SARS-Cov-2, a promising target involved in suppressing the host immune function. The authors use a combination of computational, in vitro and in vivo studies. They start with a virtual screening of the FDA approved drug library against the C-terminal domain of Nsp1, then they use biophysical assays (BLI and NanoDSF) to measure the binding affinity of the top hits emerging from the virtual screening as well as molecular dynamics simulations and MM-PBSA based free energy calculations to confirm the binding poses. Finally, luciferase-based and plaque assays were used to quantify the translation inhibition and rescue as well as infectious virus particles. One of the FDA approved drugs, Montelukast, emerges as a promising antiviral drug candidate.

Overall, the proposed drug seems to have some activity against SARS-Cov-2 (albeit the error bars reported in panel A and D of Figure 4 are quite sizeable), but there are a number of questions raised by different aspects of the pipeline.

Reviewer #3 (Public Review):

The strength of this work lies in the simplicity of the screening procedure using computational techniques resulting in the identification of an inhibitor that has the expected effects in in vitro and cellular assays. Often, this is not the case, but the current study is a good example of a successful computational screening campaign. In my opinion, one weakness of the manuscript is the lack of structural experimental data for the nsp1-montelukast complex, such as NMR or x-ray crystallography, although the authors provide a potential binding mode using computational methods. Most of the claims are justified with two minor exceptions. I think the authors conducted cellular assays rather than in vivo assays as they claim. In vivo assays would include for example mouse models, ferrets or non-human primates. This can be easily corrected and there is often confusion (in the literature) over what are cellular and in vivo assays. The authors also state that montelukast can be used to treat SARS-CoV-2 ("montelukast as a potential antiviral drug against SARS-CoV-2 infection that may help in combatting the COVID-19 pandemic"), but I don't think this statement is correct. Although the authors mention a clinical trial on montelukast during the discussion it is not obvious that the observed effects were related to inhibiting nsp1. I would rather argue that montelukast is a good starting point for the development of more potent drugs.

Reviewer #2 (Public Review):

The observation that human somatic centrioles are not molecularly rotationally symmetric, despite their structural symmetry, and that this asymmetry mediates appendage formation and ciliation is intriguing. The authors provide a fair evidence that LRRCC1, human ortholog of Vfl1, which is responsible for imparting rotational centriole asymmetry in flagellated organisms, localizes asymmetrically within mouse ependymal ciliated cells, and human RPE1. In mouse cells, LRRCC1 localizes opposite to the basal foot in ciliated centrioles. However, inconsistencies in the pattern and the levels of LRRCC1 in RPE1 across figure panels need to be clarified. The authors further demonstrate that discovered rotational asymmetry is not linked to a role in centriole duplication and there are no concerns regarding this finding.<br /> To unravel the functional role of LRRCC1 in human cells, the authors generate three CRISPR clones with decreased levels of LRRCC1. These clones have decreased rate of ciliation and perturbed ciliary signaling. As a possible cause for ciliary abnormalities, the authors suggest that LRRCC1, together with distal protein C2CD3, regulates the assembly of centriole distal ends and distal appendages. However, LRRCC1 is only partially removed, and only one of three clones forms slightly longer centrioles, despite having similar levels of LRRCC1 to other clones. A very limited analysis of centriole ultrastructure in this clone has been provided, showing one centriole pair with one longer than average and possibly structurally aberrant centriole. So, it remains unclear whether this information is relevant to the observed issues with ciliation and how lack of LRRCC1 affects centriole distal end in general. The authors also suggest that localization of distal appendage protein Cep164 is perturbed after LRRCC1 downregulation. This analysis would need to be extended to other appendage proteins.<br /> Finally, the authors propose that LRRCC1 affects localization of another distal centriole protein C2CD3. They further suggest that C2CD3 is also asymmetrically localized within centriole distal lumen, where it partially colocalizes with LRRCC1. However, the claims regarding C2CD3 asymmetry and its disorganization in 1.1 and 1.9 clones would need further evidence.

Reviewer #1 (Public Review):

In this work the authors investigate whether centrioles of the human centrosome, which are composed of 9 symmetrically arranged microtubule triplets, may display rotational asymmetry. This feature was previously described for centrioles in flagellated protists and in multi-ciliated cells, whereas human centrioles in the centrosome do not display any obvious rotational asymmetry. However, by beautiful ultra structure expansion microscopy imaging they show asymmetric localization on one side of the distal centriole lumen of the centriole protein LRRCC1, the ortholog of a protein originally shown in flagellate green algae as being asymmetrically localized. They also show that LRRCC1 affects the recruitment of another centriole protein, C2CD3, which also displays asymmetric, albeit not identical, distribution in the distal centriole lumen. In addition, using partial depletion, the authors provide data implicating LRRCC1 in proper centriole architecture, primary cilium assembly and ciliary signaling.<br /> The most important contribution of the paper is the demonstration that human centrioles are generally asymmetric, contrary to what is commonly believed, and suggesting that this is an evolutionary conserved feature. The fact that the identified asymmetrically localized protein LRRCC1 is implicated in human ciliopathy suggests that the asymmetry is functionally important. Unfortunately, in functional studies the authors achieved only partial LRRCC1 depletion and could not achieve rescue, presumably due to toxicity associated with altered LRRCC1 levels. While the observed phenotypes were relatively mild, the data using CRISPR/Cas9 genome editing as well as RNAi still support the notion that asymmetry is functionally important.<br /> Together this is a very well presented study, executed with high technical quality, that introduces rotational asymmetry as an important structural and functional feature of centrioles of the human centrosome.

Reviewer #3 (Public Review):

The authors make the important discovery that LRRCC1 localization breaks the radial symmetry of the animal centriole, preferentially associating with two consecutive triplets opposite the basal foot. LRRCC1 partially co-localizes with another centriolar protein C2CD3. Depletion of LRRCC1 altered distal appendage production and ciliary recruitment of Smoothened, but whether these requirements reflect the asymmetric nature of its centriole localization remains unclear.

One of the conclusions conflates two observations: the authors state that they "uncover the unanticipated rotational asymmetry of centrioles in the human centrosome and show that this property is connected to the assembly and function of primary cilia." The rotational asymmetry is an advance, but whether it is this asymmetry that is important to assembly and function of cilia or some other aspect of LRRCC1 function remains unclear. Said another way, while LRRCC1 and C2CD3 are localized asymmetrically at centrioles, there is no test of whether their asymmetry directly contributes to the assembly and function of primary cilia.

Reviewer #1 (Public Review):

In this study, Beon et al., show that the Inositol polyphosphate multikinase enzyme (IPMK) interacts with several subunits of the SWI/SNF chromatin remodeling complex. They describe a direct interaction between the SMARCB1 (BAF47) subunit and IPMK and determine the regions of each protein required for interaction. Using ChIP-seq in presence or absence of IPMK siRNA silencing they show that IPMK modulates BRG1 occupancy primarily at the -1 and/or +1 nucleosome at the transcription start site. BRG1 occupancy is preferentially affected at promoters with bivalent chromatin modifications and its diminished occupancy regulates gene expression. The authors show convincing data that IPMK interacts with SWI/SNF and modulates is genomic occupancy in embryonic stem (ES) cells to regulate gene expression. However, the study does not address if these effects regulate ES cell differentiation nor whether they involve IPMK enzymatic activity.

Reviewer #2 (Public Review):

SWI/SNF subunits were identified as IPMK interacting proteins in two unbiased screening assays (yeast two hybrid with IPMK as bait and a human cDNA library as prey as well as in vivo proximity-labeling). The interactions were further characterized in mammalian cells using over-expressed tagged proteins and endogenous proteins by immunoprecipitations. Direct interactions were elucidated using baculovirus-purified proteins and interaction domains identified by deletion studies. Overall the protein-protein interaction studies are very convincing with the exception of one endogenous co-immunoprecipitation. The cut and run as well as the ATAC-seq data also look strong. However, important experimental and analysis details are missing regarding. While the emphasis of the chromatin occupancy and accessibility studies is on the regulation of BRG1 by IPMK, the manuscript loses focus by performing gene expression profiling in Smarcb1 and not BRG1 depleted cells. Suggestions for increasing the focus and broader appeal of the work include providing a more extensive integration of chromatin occupancy and chromatin accessibility data with the effects of either BRG1 or Smarcb1 knockdown on gene expression.

In its current state, the findings may not be of broad interest. Experiments to link the molecular studies of IPMK depletion with biological effects of IPMK depletion on aspects of embryonic stem cell biology would increase the impact of the work and broaden the interest level to other fields.

Reviewer #3 (Public Review):

The authors explored in mammalian cells the linkage between inositol polyphosphates, chromatin remodeling, and transcription regulation.

They first used a combination of experimental approaches including yeast two-hybrid screening, in vivo proximity labeling, in vitro binding assays and co-immunoprecipitation experiments to show that IMPK interacts with several subunits of the mammalian SWI/SNF complex. Altogether, these experiments provide strong evidence that IMPK and SWI/SNF complex(es) interact in vivo.

They next used CUT&RUN and ATAC-seq to probe the importance of the interaction between SWI/SNF and IMPK for chromatin remodeling at transcription cis-regulatory elements.

A major concern of the BRG1 CUT&RUN experiments realized in mouse ES cells is that the authors did not identify enhancers as regions of enrichment for BRG1 : enhancers are expected to be present in the intergenic fraction of the genome, which in the manuscript was identified as lacking BRG1 enrichment (Figure 4D,E). Previous publications and their associated ChIP-seq datasets for BRG1 in mouse ES cells and other cell types all revealed a strong BRG1 enrichment at enhancer elements (PMID: 28945250, PMID: 25803486, PMID: 26814966). This discrepancy suggests that something might have gone wrong with the BRG1 CUT&RUN experiments, and thus it is very important that the authors clarify this point.

Another concern is that the impact of IPMK knockdown on BRG1 enrichment appears to be very mild in Fig. 4B,C,F,G. Interpretation of low amplitude changes in ChIP-seq or CUT&RUN signal is always a difficult task because non-specific variations in signal might make strong contributions to heat-map and average profile analyses. Further analysis of the data should be done to validate these low amplitude changes.

Reviewer #1 (Public Review): 

The present work by Phillips et al., builds on a previously published (eLife 2019, 8:e41555) computational model that showed how rhythmicity and the amplitude of respiratory oscillations involve distinct biophysical mechanisms. In particular, the model predicts that respiratory rhythm can be independent of calcium-activated non-selective cation current activation, and that this determines population activity amplitude. In contrast, rhythm depends on sodium currents in a subpopulation of cells forming a preBötC rhythmogenic kernel. The past model proposed by Phillips et al., (2019) consistently reproduced some previously published experimental studies. 

The experimental data obtained in this current work systematically demonstrate that some of the simulations and predictions generated from their computational model are accurate, thereby illustrating the robustness of their computational model. 

Strengths: 

Both the computational model and empirical data provided in this work further foster our understanding on how the preBötC generates (respiratory/inspiratory) rhythmogenesis and highlights the existence of distinct biophysical mechanisms involved in rhythmicity and the amplitude of respiratory oscillations. Collectively, this work is of great interest to the respiratory neuroscientist community. 

Weaknesses:

Whereas the major claims of this work are supported by solid experimental data, the manuscript is written in a highly technical manner that is not comprehensible for scientists not familiar with computational modeling and electrophysiology. It would be desirable that the authors could make the text more accessible to a larger audience.

Reviewer #2 (Public Review): 

In this manuscript, Phillips et al. address the relevance of the persistent inward conductances, INaP and ICAN, for inspiratory rhythm and pattern generation. The authors previously developed a computational model of the inspiratory rhythm generator, the preBötzinger Complex (preBötC), that relied on INaP for rhythm generation and ICAN for pattern generation. Here, they perform experiments designed to test certain predictions of their model using thin rhythmic medullary slices from triple transgenic mice where both tdTomato and ChR2-EYFP are expressed in glutamatergic VGLUT2-expressing neurons. The authors show that pharmacological blockade of INaP leads to dose-dependent decreases in burst frequency and amplitude under baseline conditions and at varying levels of tonic optogenetic excitation with high concentrations of the blocker preventing rhythmic bursting even at high laser powers. Pharmacological blockade of ICAN reduces amplitude, but does not significantly affect frequency at baseline and causes an increase in frequency when laser power is increased. The authors make the claim that these data support their model and the hypothesis that INaP is essential for preBötC rhythmogenesis and ICAN is essential for determining burst amplitude, but is dispensable for rhythm generation. 

The strengths of the manuscript are that the computational model is revised to include a biophysical model for channelrhodopsin and that the modeling and experiments support the proposed role for ICAN in burst generation. The prediction of an increase in frequency with increased tonic excitation when ICAN is blocked is of particular interest. 

Despite these strengths, a number of major issues significantly weaken the manuscript and limit its impact in advancing understanding of rhythm and pattern generation in preBötC. 

1) Optogenetic stimulation. The authors use an optogenetic approach that may be more complex than assumed and that is not adequately validated in their model. The transgenic mouse used expresses both tdTomato and ChR2-EYFP in all glutamatergic VGLUT2-expressing neurons. The authors assume that bilateral illumination over preBötC enables depolarization specifically in the preBötC excitatory population. However, ChR2 will be expressed in all glutamatergic neurons, so the illumination may depolarize terminals or fibers of passage from glutamatergic neurons outside the preBötC (even those whose somata were removed in slicing). Illumination of non-rhythmogenic preBötC neurons may affect interpretation of their results and congruence with their model, which only contains the preBötC rhythmogenic population. Furthermore, ChR2-induced depolarization may interact unexpectedly with other membrane properties and conductances. Two examples that indicate that the optogenetic stimulation protocol may not be straightforward is the 1-3 minute inhibition of rhythmicity following illumination (p 8, line 20-22, Figure 3B) and what appears to be a hyperpolarization following 5 mW illumination in their whole cell patch clamp recordings (Fig 2C). While the voltage dependence of ChR2 in their model is presented, whether these other phenomena are also reproduced in their model is not demonstrated, calling into question how to interpret their comparisons of experimental and model results. 

2) Challenges to the INaP hypothesis in published results. The biggest issue with the manuscript is its central hypothesis that INaP is essential for rhythmogenesis. This hypothesis has faced considerable scrutiny, and a number of papers appear to invalidate a necessary role for INaP in preBötC rhythmogenesis. The authors mention these other results superficially in the Discussion, but do not grapple with their clear challenges to the INaP hypothesis. Pace et al. (2007) showed that bilateral microinjection of riluzole or low concentrations of TTX into preBötC failed to stop the rhythm and that the pharmacological effects of these blockers could be explained by their effects on raphe excitability, which provides tonic excitatory drive to the preBötC. The authors propose that these conflicting results can be explained by differences in slice thickness and incomplete pharmacological penetration; however, the Pace paper specifically addressed this issue by microinjecting the drugs 100 um below the surface. Further, the raphe microinjection provide an alternative experimentally-validated explanation for many prior and current pharmacological experiments involving INaP blockers. All blockers were bath-applied here, and these concerns were not addressed experimentally. Finally, off-target effects of INaP blockers, particularly at the higher concentrations, were also not addressed. In addition to INaP blockers, other published results show that rhythmicity can occur without high frequency bursts necessary for INaP activation, and pharmacology experiments in situ also suggest that rhythmicity can persist in more intact networks without INaP. These issues are discussed but not addressed experimentally. Thus, the substantial body of experimental work that is inconsistent with the INaP hypothesis remains relevant. 

3) Model limitations. Experimental confirmation of a limited set of predictions of a reduced model does not strongly support a particular model mechanism if the model does not include known conductances/properties of the biological system and does not reproduce other experimentally observed phenomena. Without including burst-terminating conductances, physiological connectivity and synaptic properties, and perhaps other preBötC populations, e.g., inhibitory neurons, the experimental results may not uniquely validate the model. Additionally, the model should be capable of reproducing a variety of experimentally observed preBötC phenomenology besides those directly related to INaP and ICAN. Without such constraints, the model could easily be tuned (and is in fact tuned in this manuscript) to reproduce selected results, severely limiting the validity and generalizability of the model and its mechanisms. 

4) Statistical comparisons. Statistical comparisons are relatively limited in this manuscript. Methods mention Student's t-test or the Wilcoxon signed rank test, but it appears that some of the data, e.g., frequency/amplitude dose dependent curves, downward shifts of frequency or amplitude in drug, and comparisons between model and experiment, would require parametric or non-parametric multiple comparison tests, such as ANOVA or Kolmogorov-Smirnov. Without such comparisons, qualitative descriptions may mask non-significant variations or statistically significant differences may be missed.

Reviewer #3 (Public Review): 

In "Predictions and experimental tests of a new biophysical model of the mammalian respiratory oscillator" the authors test three hypotheses: 1) INap and ICAN blockade alter network excitability in the preBötC, the region of the brainstem that generates inspiratory breathing rhythm; 2) that INaP is essential for preBötC rhythmogenesis; 3) ICAN is essential for generating the amplitude of rhythmic output but not rhythm generation. They test these hypotheses using optogenetic manipulation of local preBötC excitability and the use of pharmacologic blockade of INaP and ICAN. 

The manuscript is well-written and clear. The experiments are appropriate to test the hypotheses and the data is convincing. This manuscript is significant because it provides substantive evidence for the role of INaP in modulating breathing frequency and ICAN in altering amplitude with some interesting boundary conditions when ICAN and INaP are selectively blocked. Of particular value is the addition of a channelrhodopsin current (based on a Markov formalism) to the authors' previously published model. 

The authors provide strong evidence testing their hypotheses and showing the importance of both INaP and ICAN for the generation of reliable breathing rhythm. 

The results presented here provide strong evidence for separate roles for INaP and ICAN in modulating frequency and amplitude of central respiratory drive.

Reviewer #1 (Public Review): 

In this Tools and Resources manuscript, Gargareta/Reuschenbach/Siems/Sun and colleagues use quantitative proteomics to analyze myelin from human white matter. This new dataset was rigorously generated - post-mortem variability was well-controlled, the unlabeled approach allowed for abundance analysis, and two runs further helped control for mass spec variability. Following the generation of this new human myelin proteome dataset, the authors compared it to their previously generated mouse myelin proteome dataset, and discovered a number of surprising species-specific results that were properly followed up in primary tissue, including the presence of previously thought to be PNS myelin specific protein PMP2 in human CNS myelin. Given that the myelin field predominately uses mouse models to study development and disease/injury states, being aware of these key differences is critically important. The authors further compare their proteomics datasets to several existing transcriptomics datasets for both species, and find that these are well-correlated. In all, this is an excellent new resource for the myelin community as well as for anyone interested in how myelinating oligodendrocytes might contribute to human disease. The following are minor suggestions that the authors could consider: 

• The text regarding Figure 3, Supplement 2 could have a more detailed description to underscore the significance of this data.

• Can the authors clarify brain region(s) in the text/methods beyond "white matter?"

• This resource would be of even greater utility to the community if there was any way to generate a searchable database, similar to https://www.brainrnaseq.org/. I realize there is a lot that goes into not only generating such a website but also maintaining it and that this may not be immediately possible.

Reviewer #2 (Public Review): 

This manuscript covers a worthy topic to understand mouse versus the human myelin proteome. These data have far reaching implications for translating studies in preclinical mouse models to humans. This study takes a deep dive into understanding differences between mouse and human myelin using both quantitative proteomics with comparison to published scRNA-seq datasets. Important comparisons are presented with relevant proteins known to drive myelin formation. Overall, this is a comprehensive study that will provide an important resource to the field.

Reviewer #3 (Public Review): 

In "Conservation and divergence of myelin proteome and oligodendrocyte transcriptome profiles between humans and mice" the authors profile the molecular composition of myelin in both humans and mice using mass spectrometry and complement this by analysing expression of mRNA in myelinating oligodendrocytes. The aim of the study is to investigate to what extent the molecular composition and general profile of myelin varies between species. To date, the mouse model has been the pre-dominant animal model for the study of myelin, an essentially vertebrate-specific feature that has allowed the evolution of large and complex nervous systems. Nonetheless, mice and humans diverged over 85 million years ago and it has remained unclear to what extent the molecular composition and in turn regulation of myelin and myelinating oligodendrocytes is similar or distinct between species. This study uses mass spectrometry to define the proteome of human myelin and compares these new data with previous profiling of myelin in mice. The very broad and deep analysis in the manuscript shows that the relative abundance of the small number of factors that comprise the majority of total myelin protein is very similar between mice and humans. In contrast, however, the study unmasks interesting differences in many less abundant proteins, with several found only in human myelin and others only in the mouse. The authors support these findings in follow-up studies that assess protein presence directly in tissue, and also by comparing the mRNA profile of human and mouse myelinating oligodendrocytes directly, finding that by and large the findings made at the protein level are predictable from their mRNA expression. This set of data will be enormously useful to the community, with a wealth of information available that is not feasible to convey in one manuscript narrative. It is also not feasible to assess the functional relevance of the species-specific aspects of mouse or human myelin in this initial study, but those will be very important avenues for future exploration. I expect this manuscript to greatly inform ongoing studies of myelin in both health and disease.

Reviewer #1 (Public Review): 

Hu and colleagues employ computed-tomography methods and provide a detailed description of and inferences about the dental system in three early-diverging ceratopsian dinosaur genera represented by rare specimens from China. Their study identifies nuanced tooth replacement rates and patterns. Furthermore, combined with the analysis of dental wear patterns, their study not only elucidates ontogenetic aspects of these early ceratopsians but also explores the implication of such patterns for dietary adaptations among these taxa. The manuscript, therefore, provides unique insights into the anatomical and ecological contexts of ceratopsians in such deep time. 

The manuscript is rich in data that are summarized in multiple tables and figures. It is also well-written and easy to follow. The inference and conclusions made are also overall well supported by the data presented. 

The only main comment I have concerns the inference made about the dietary adaptation of Yinlong, which is inferred to be characterized by "feeding strategies other than only grinding food with their teeth." I think that this could be expanded a bit more to incorporate dietary breadth as an additional possible explanation, particularly given the lack of conclusive evidence for the predominance of a single plant species. As it stands, the inference (made across lines 475 through 485) may only imply processing the same food resource using non-chewing methods (e.g., gastroliths to triturate fern). Could the incorporation of other, less abrasive plat foods--in addition to the fibrous ferns--in the diet of Yinlong be a possible, additional explanation for the relatively slow tooth replacement and lack of a heavy tooth wear from chewing-related stress?

Reviewer #2 (Public Review): 

The authors of the present work aimed to describe tooth replacement in early ceratopsian species from the Lower Jurassic of China, and with this novel information, discuss new hypotheses of successive changes in jaw evolution that led to the highly specialized replacement and jaw function of derived ceratopsids. Major strengths of this study include not only the use of microCT-scans and 3D reconstructions to address tooth replacement in three different species of early ceratopsians (Yinlong, Hualianceratops, and Chaoyangsaurus), but also the observation of wear development, pulp cavity development, zahnreihen, and z-spacing and replacement rate to compare between taxa and address the succession of mandibular and replacement changes in the phylogeny of ceratopsian dinosaurs. The aims were achieved and the conclusions are strongly supported by the evidence discussed and the cited bibliography. Figures are clear and captions are concise. The presented information gives evidence for the comparison and discussion of the order of acquisition of different craniomandibular adaptations that lead to a specialized herbivorous diet, useful not only for ceratopsians and ornithischians, but also for other lineages of dinosaurs in the Mesozoic, and further for comparing with extant and extinct lineages of mammals. Dinosaurs not only were fantastic creatures from the past but also achieved different morphologic, physiologic, and behavioral traits unknown to any other creature, even mammals. For ceratopsians, the appearance of dental batteries corresponds to a unique trait only functionally similar to that in hadrosaurs and some sauropods, and understanding the steps that led to that specialized structure allows us to also understand the drivers that later guided their diversification during the Late Cretaceous.

Reviewer #3 (Public Review): 

The major strengths of the paper are its thorough level of detail, rich dataset, and easy readability. The figures are excellent and clear. 

One shortcoming of the paper is the lack of measurements -- a table of measurement for each functional and replacement tooth's length, mesiodistal width, and linguolabial width should be provided. 

Unfortunately the manuscript is not publishable in its current form because the conclusions are not testable based on the limited data provided. The authors stated "All data generated or analysed during this study are included in the manuscript and supporting file." This is not true. Only the 3D models derived from segmentations are provided, not the raw scans. Segmentation-derived models are interpretations, akin to publishing a drawing of a fossil instead of a photograph, which is not generally acceptable under today's publishing standards (drawings can be published alongside photographs). Please upload the raw scans to an appropriate repository such as Morphosource, Dryad, or Morphobank. Scans can be cropped to the dentigerous regions only, so long as scaling information is preserved.

Reviewer #1 (Public Review): 

The authors are providing new important data through which to discuss the evolution of the scapula and coracoid and how it relates to the evolution of the modern avian flight stroke. 

The strength lies in the excellent new CT data. Weaknesses include inconsistency in terminology and a lack of recognition for previous work done. 

Information regarding the precise 3D morphology of the scapulocoracoid in Sapeornis is a major step in the direction of better understanding this important transformation in birds. The data once publically available can be used to further explore flight capabilities in this important taxon.

Reviewer #2 (Public Review): 

The manuscript by Shiying Wang and co-authors entitled "Digital restoration of the pectoral girdles of two Early Cretaceous birds, and implications for early flight evolution" constitutes a wonderful and highly detailed study on two important fossil birds. For the first time a 3D reconstruction of their scapular girdle is provided, offering novel anatomical data which clarify the sequence of morphological features in the line to crown-birds. I have enjoyed reading this contribution. 

There are some minor aspects that the authors will be able to solve. Some other points concern the phylogenetic interpretations (mainly concerning the relationships of Rahonavis), for which I recommend modifications, that the authors should feel free to follow or not. 

I disagree with some topical aspects on the morphology of the coracoid of Sapeornis and the inferred function of such features. Nevertheless, even if the authors accept my thoughts on this issue, the main interpretations offered in the ms will not change, and only the origin of some adaptations (e.g., protactor vs elevator function of the m. supracoracoideus) would need to be revised. However, even if the authors do not accept my interpretations, this is a valuable "step forward" paper on the anatomy and evolution of early birds.

Reviewer #1 (Public Review):

Single point measurements of pulse wave velocity may be a way to bring the measurement of arterial stiffness into daily clinical practice. However, the advantages and drawbacks of such a method need to be carefully investigated. This short communication aims to identify shortcomings of the so-called "ARCSolver" algorithm for pulse wave velocity estimation. Unfortunately, this manuscript does not bring many additional insights to the existing knowledge. We will assess their arguments in the following discussion:

The authors argue that the PWV measurements did not meet the expectations by master athletes, i.e. were higher than athletes and study personal expected. However, no other assessment of vascular properties in these subjects is presented, which could back up the argument. Thus, no scientifically solid conclusion can be drawn from this data. The expectation that older athletes will have a lower PWV compared to sedentary subjects is based on a paper from Vaitkevicius et al. In that study, only 14 senior athletes were included, and PWV measurements were a mixture of two methods. More recent studies in marathon runners show that athletes had a significant higher PWV than controls (Vlachopoulos et al, doi: 10.1038/ajh.2010.99), had increased coronary artery plaque volume compared to controls (Schwartz et al, PMID: 30323509) and that conventional cardiovascular risk stratification using the Framingham Risk score underestimates the CAC burden in presumably healthy marathon runners (Möhlenkamp et al, doi: 10.1093/eurheartj/ehn163).

The authors also argue that it unexpected that PWV values from master athletes lie on the same regression line as PWV values from participants of a head down tilt bed rest study called AGBRESA. When we understand correctly, the authors would expect that the PWV values from the master athletes should be comparably lower when trajected to the age of the bed rest study participants. However, when looking at the inclusion criteria for the AGBRESA study, only very healthy subjects were included. Furthermore, study participants were rather young, where no manifest arterial changes can be expected. Thus, when assuming that master athletes are also relatively healthy, it would actually be fully logical that they lie on the same regression line as the bed rest study participants.

Within the subjects of the bed rest study, a comparison to a PWV estimated from the pulse arrival time at the thigh (corrected by isovolumic contraction time) was performed. It was found that the difference between this PWV and the ARCSolver PWV correlates with the age of the subjects. This is an interesting finding. However, this does not show that the one or the other method is not reliable. It is already well known from literature, that aortic (invasive) PWV and carotid-femoral PWV have different trajectories over age, thus values are not directly comparable. While the ARCSolver PWV was developed to estimate aortic PWV, the PAT-based PWV in this study rather mimics a carotid-femoral PWV. In consequence, a diverging prolongation over age is expected.

In another experiment, the authors measured ARCSolver PWV repeatedly in the same subject, but changed the age of the subject in the software. Then they concluded that the changes in the resulting PWV almost exclusively depend on the entered age. However, this is not at all surprising but is exactly what should be the result of this experiment. It is known from previous publications that ARCSolver PWV is estimated from age, systolic blood pressure and waveform information. Since only age has been altered as an input in this experiment, the change in output can only depend on age as well.

In conclusion, the role of different methods for PWV estimation still needs to be found. While some studies have found a prognostic value of estimated PWV, other studies criticise estimation methods to be too simplistic and not capable of assessing vascular aging as detailed as needed. Thus, further studies are needed to bring more clarification.

Reviewer #2 (Public Review):

This short report reveals a profoundly important issue relating to at least one device that claims to estimate pulse wave velocity (PWV) from a single arm cuff recording. The authors reveal a near-perfect correlation between age and estimated PWV, strongly implying that the device makes use of a statistical model that is almost entirely dependent on entered age. This finding is not surprising, given similar high (and unbelievable) correlations with age reported in prior work. The authors also showed that a factor other than age (long term exercise) that is expected to have an impact on actual PWV, had no impact on estimated PWV. Such a device could therefore lead to false conclusions in scientific studies. The authors rightly conclude that PWV estimated by such a device is no substitute for directly measured PWV via, e.g. the carotid-femoral technique.

Reviewer #1 (Public Review):

While it is known for a long time that transcription factors control the identity of individual neuron types by acting either as activators or repressors of gene expression, very little is known about the mechanisms controlling chromatin accessibility in post-mitotic neurons. Most studies on this topic are done in vitro due to the low abundance of many post-mitotic neuron types in vivo. However, the current study bypasses this limitation by employing cutting-edge methods (e.g., single cell ATAC-Seq, RNA-Seq, ChIPmentation) to establish maps of accessible chromatin and cis-regulatory elements in the context of developing, postmitotic serotonergic (5-HT) neurons of the mouse hindbrain. Major conclusions include:

1. Diverse, single-cell chromatin landscapes are stablished early in postmitotic Pet1 neurons and likely account for the transcriptomic heterogeneity of adult 5-HT neuron subtypes.

2. By comparing chromatin accessibility across different developmental stages, they found that chromatin remodeling is highly dynamic following cell cycle exit but later stabilizes.

3. The terminal selector-type transcription factor Pet1 reorganizes chromatin accessibility of genes necessary for serotonin (5-HT) biosynthesis.

4. The early euchromatin landscape of 5-HT neurons is dynamic and requires continuous Pet1 activity for maintenance of chromatin accessibility.

5. The terminal selector-type transcription factor Lmx1b has a broader impact than Pet1 on serotonin (5-HT) neuron chromatin.

The conclusions are supported by the experimental evidence.

Strengths:

Technically, this is a tour-de-force effort. The study employs cutting-edge molecular, genetic and biochemical methods - in vivo - to study the molecular mechanisms underlying serotonergic neuron maturation in wild-type and conditional mouse mutants for Pet1 and Lmx1b. Essential controls are included and the authors provide all necessary information to help the reader understand how the analysis of all datasets was performed.

Conceptually, this study advances the field by shedding light into the poorly understood mechanisms governing chromatin accessibility in maturing neuron types. First, the authors found that heterogenous chromatin landscapes are established early during development in Pet-1-neurons, possibly setting the stage for the generation of distinct subtypes of 5-HT neurons. Second, they propose a new function for terminal selector-type transcription factors. That is, Pet1 and Lmx1b control 5-HT neuron maturation not only through sequence-specific activation of terminal effector genes, but also by reorganizing accessible chromatin at cis-regulatory regions in 5-HT neurons.

Weaknesses:

The mechanistic details of how Pet1 and Lmx1b open chromatin remain elusive.

Reviewer #3 (Public Review):

In this manuscript Zhang et al. perform an extensive analysis on the chromatin accessibility state of serotonergic neurons at different stages in development, in wild type and Pet1 and Lmx1b mutant backgrounds, two transcription factors directly involved in the terminal differentiation of serotonergic neurons.

Authors also present data on epigenetic marks on these neurons and single cell ATACseq and scRNAseq data for the wildtype e14.5 embryonic stage.

There is an impressive amount of data that nicely shows the dynamics of chromatin accessibility along serotonergic neuron maturation and the diversity of chromatin states in different 5HT neuron subtypes.

Most of the analysis reinforces known roles for Pet1 and Lmx1b however, it does not significantly increase our understanding of important questions that remain unknown for serotonergic specification and could be of broad intestest: 1) how is sustained expression of Pet1 and Lmx1b combined with additional mechanism to regulate dynamic chromatin landscapes and gene expression during serotonergic maturation?; 2) Are Pet1 and Lmx1b TFs involved in subtype diversity of serotonergic neurons? 3) If so, how is broad Pet1 and Lmx1b expression translated into gene specific defects in 5HT neuron subtypes? 4) Is Pet1 and Lmx1b binding dynamic along postmitotic serotonergic development? 5) How do Pet1 and Lmx1b mediate repression of gene expression?

Addressing some of these questions might require additional experimental approaches and experiments.

Reviewer #2 (Public Review):

Zhang et al examined regulatory regions controlling expression of genes in embryonic serotonergic neurons. They mapped accessible chromatin regions and gene expression using bulk and single cell analysis, performed chromatin immunoprecipitation studies of Pet-1 and Lmx1b transcription factors and investigated effects of Pet1 and Lmx1b loss of function on chromatin accessibility. The study generated a useful resource of genomic data and provided new insights into the dynamic regulation of accessible chromatin regions in postmitotic neurons.

Despite the wealth of new genomic data, the focus of the manuscript is somewhat elusive. The goal seems to be to study the mechanisms that control chromatin accessibility in postmitotic serotonergic neurons during their subtype diversification. The authors provide in depth analysis of Lmx1b and Pet1 binding and effects of knockouts of these factors on chromatin accessibility. The authors also rely on conditional deletion of these TFs to address whether the factors are required to "maintain" TACs. Unfortunately, analysis of these conditional mice is limited to one timepoint, which complicates interpretation of the data. It is also unclear why the authors do not show more global analysis of gene expression changes in the pet1/Lmx1b DKO.

In summary, while the manuscript provides important new data, the model how serotonergic neuron specification is controlled by Pet1 and Lmx1b remains unclear. Most importantly, the interesting single cell data are not well integrated with transcription factor function and with temporal changes in chromatin accessibility. Overall, the manuscript will benefit from clearer writing, explicit stating what hypotheses are being tested and better explanation of the rationale for the performed experiments.

Reviewer #2 (Public Review):

Identifying and quantifying the tumor specific T cell response continues to be an area of great interest in the context of cancer immunotherapy. The presence of sets of T cells with similar sequence (clusters) is becoming increasingly accepted as a key feature of the antigen-specific T cell response. In this study the authors explore this hypothesis in the context of melanoma. The study uses the bioinformatics tool ALICE which identifies TCR clusters, taking into account "background" clustering which might be expected by chance.

The study reports three key findings: 1. the number of clusters in TILs increases following anti-PD1 immunotherapy. 2. the number of clusters is enriched in the CD39+PD1+ positive TIL fraction. 3. the number of clusters can be used as an indication of the success of in vitro expansions of tumor specific T cells.<br /> Overall, the data provided are convincing, although the number of samples studied is small. The explanation of the figures is occasionally hard to follow, which in its current form lessens the impact of the paper. But the study will certainly be of interest to all those interested in characterising the tumor-specific T cell response in humans, and could form the basis for further more extensive studies to validate the results, and apply them to well-characterised clinical cohorts.

Reviewer #3 (Public Review):

Goncharov et al provide a clear example of the importance of immune receptor repertoire profiling to identify responding tumor-specific T cells. They show that a previously proposed algorithm (ALICE) can be a useful tool to characterize tumor-specific T cell enrichment in grafts and to optimize TIL cultures. These results have the potential to accelerate clinical development of adoptive T cell transfer techniques and are of interest for the community.

The paper is concise and to the point. The conclusions are well supported by the data and the analyses. Some additional explanations regarding the computational analysis can help readability and clarity.

Reviewer #1 (Public Review):

The authors investigate a clustering-based method to find reactive T cells based on their TCR (T cell receptor) sequences following ACT (Adoptive T cell transfer). This method, which was previously implemented as ALICE, find reactive T cell clones in samples by looking for overrepresented clusters of T cells with similar TCR sequences.

By applying the method on published data from Melanoma patients, the authors show an increase in the number of clusters following anti-PD1 immunotherapy. They also find in those clusters many TCRs known to be reactive to melanoma antigens. Clusters are also found in CD39+PD1+ activated T cells.

Overall, the paper shows strong indications that clusters are indeed enriched for tumor reactive TCRs. The overall number of reactive TCRs in the clusters, on the other hand, is not known (and hard to estimate). Specifically, it is not clear how many of the TCRs in the clusters found using this method are indeed reactive against the tumor cells. However, the ones found are excellent candidates for functional assays that determine reactivity. The functional analysis presented in the paper, which involved sorting on CD137, doesn't link the TCRs in the clusters with activation very strongly.

The paper makes a strong case to the usefulness of cluster-based analysis for measuring tumor related response and finding possible reactive TCRs. However, stronger functional validation methods are needed to assess the quality of the TCRs found in the clusters. Further work would investigate this relation in more depth, mainly to pinpoint and improve the sensitivity and accuracy of the inferred tumor related clones.

Reviewer #1 (Public Review):

In this manuscript, the authors introduce a new piece of software, BehaviorDEPOT, that aims to serve as an open source classifier in service of standard lab-based behavioral assays. The key arguments the authors make are that 1) the open source code allows for freely available access, 2) the code doesn't require any coding knowledge to build new classifiers, 3) it is generalizable to other behaviors than freezing and other species (although this latter point is not shown), 4) that it uses posture-based tracking that allows for higher resolution than centroid-based methods, and 5) that it is possible to isolate features used in the classifiers. While these aims are laudable, and the software is indeed relatively easy to use, I am not convinced that the method represents a large conceptual advance or would be highly used outside the rodent freezing community.

Major points:

1) I'm not convinced over one of the key arguments the authors make - that the limb tracking produces qualitatively/quantitatively better results than centroid/orientation tracking alone for the tasks they measure. For example, angular velocities could be used to identify head movements. It would be good to test this with their data (could you build a classifier using only the position/velocity/angular velocities of the main axis of the body?

2) This brings me to the point that the previous state-of-the-art open-source methodology, JAABA, is barely mentioned, and I think that a more direct comparison is warranted, especially since this method has been widely used/cited and is also aimed at a not-coding audience.

3) Remaining on JAABA: while the authors' classification approach appeared to depend mostly on a relatively small number of features, JAABA uses boosting to build a very good classifier out of many not-so-good classifiers. This approach is well-worn in machine learning and has been used to good effect in high-throughput behavioral data. I would like the authors to comment on why they decided on the classification strategy they have.

4) I would also like more details on the classifiers the authors used. There is some detail in the main text, but a specific section in the Methods section is warranted, I believe, for transparency. The same goes for all of the DLC post-processing steps.

5) It would be good for the authors to compare the Inter-Rater Module to the methods described in the MARS paper (reference 12 here).

6) More quantitative discussion about the effect of tracking errors on the classifier would be ideal. No tracking is perfect, so an end-user will need to know "how good" they need to get the tracking to get the results presented here.

Reviewer #2 (Public Review):

BehaviorDEPOT is a Matlab-based user interface aimed at helping users interact with animal pose data without significant coding experience. It is composed of several tools for analysis of animal tracking data, as well as a data collection module that can interface via Arduino to control experimental hardware. The data analysis tools are designed for post-processing of DeepLabCut pose estimates and manual pose annotations, and includes four modules: 1) a Data Exploration module for visualizing spatiotemporal features computed from animal pose (such as velocity and acceleration), 2) a Classifier Optimization module for creating hand-fit classifiers to detect behaviors by applying windowing to spatiotemporal features, 3) a Validation module for evaluating performance of classifiers, and 4) an Inter-Rater Agreement module for comparing annotations by different individuals.

A strength of BehaviorDEPOT is its combination of many broadly useful data visualization and evaluation modules within a single interface. The four experimental use cases in the paper nicely showcase various features of the tool, working the user from the simplest example (detecting optogenetically induced freezing) to a more sophisticated decision-making example in which BehaviorDEPOT is used to segment behavioral recordings into trials, and within trials to count head turns per trial to detect deliberative behavior (vicarious trial and error, or VTE.) The authors also demonstrate the application of their software using several different animal pose formats (including from 4 to 9 tracked body parts) from multiple camera types and framerates.

One point that confused me when reading the paper was whether BehaviorDEPOT was using a single, fixed freezing classifier, or whether the freezing classifier was being tuned to each new setting (the latter is the case.) The abstract, introduction, and "Development of the BehaviorDEPOT Freezing Classifier" sections all make the freezing classifier sound like a fixed object that can be run "out-of-the-box" on any dataset. However, the subsequent "Analysis Module" section says it implements "hard-coded classifiers with adjustable parameters", which makes it clear that the freezing classifier is not a fixed object, but rather it has a set of parameters that can (must?) be tuned by the user to achieve desired performance. It is important to note that the freezing classifier performances reported in the paper should therefore be read with the understanding that these values are specific to the particular parameter configuration found (rather than reflecting performance a user could get out of the box.)

This points to a central component of BehaviorDEPOT's design that makes its classifiers different from those produced by previously published behavior detection software such as JAABA or SimBA. So far as I can tell, BehaviorDEPOT includes no automated classifier fitting, instead relying on the users to come up with which features to use and which thresholds to assign to those features. Given that the classifier optimization module still requires manual annotations (to calculate classifier performance, Fig 7A), I'm unsure whether hand selection of features offers any kind of advantage over a standard supervised classifier training approach. That doesn't mean an advantage doesn't exist- maybe the hand-fit classifiers require less annotation data than a supervised classifier, or maybe humans are better at picking "appropriate" features based on their understanding of the behavior they want to study.

There is something to be said for helping users hand-create behavior classifiers: it's easier to interpret the output of those classifiers, and they could prove easier to fine-tune to fix performance when given out-of-sample data. Still, I think it's a major shortcoming that BehaviorDEPOT only allows users to use up to two parameters to create behavior classifiers, and cannot create thresholds that depend on linear or nonlinear combinations of parameters (eg, Figure 6D indicates that the best classifier would take a weighted sum of head velocity and change in head angle.) Because of these limitations on classifier complexity, I worry that it will be difficult to use BehaviorDEPOT to detect many more complex behaviors.

Finally, I have some concerns about how performance of classifiers is reported. For example, the authors describe "validation" set of videos used to assess freezing classifier performance, but they are very vague about the detector was trained in the first place, stating "we empirically determined that thresholding the velocity of a weighted average of 3-6 body parts ... and the angle of head movements produced the best-performing freezing classifier." What videos were used to come to this conclusion? It is imperative that when performance values are reported in the paper, they are calculated on a separate set of validation videos, ideally from different animals, that were *never referenced* while setting the parameters of the classifier. Otherwise, there is a substantial risk of overfitting, leading to overestimation of classifier performance. Similarly, Figure 7 shows the manual fitting of classifiers to rat and mouse data; the fitting process in 7A is shown to include updating parameters and recalculating performance iteratively. This approach is fine, however I want to confirm that the classifier performances in panels 7F-G were computed on videos not used during fitting.

Overall, I like the user-friendly interface of this software, its interaction with experimental hardware, and its support for hand-crafted behavior classification. However, I feel that more work could be done to support incorporation of additional features and feature combinations as classifier input- it would be great if BehaviorDEPOT could at least partially automate the classifier fitting process, eg by automatically fitting thresholds to user-selected features, or by suggesting features that are most correlated with a user's provided annotations. Finally, the validation of classifier performance should be addressed.

Reviewer #3 (Public Review):

There is a need for standardized pipelines that allow for repeatable robust analysis of behavioral data, and this toolkit provides several helpful modules that researchers will find useful. There are, however, several weaknesses in the current presentation of this work.

It is unclear what the major advance is that sets BehaviorDEPOT apart from other tools mentioned (ezTrack, JAABA, SimBA, MARS, DeepEthogram, etc). A comparison against other commonly used classifiers would speak to the motivation for BehaviorDEPOT - especially if this software is simpler to use and equally efficient at classification. While the idea might be that joint-level tracking should simplify the classification process, the number of markers used in some of the examples is limited to small regions on the body and might not justify using these markers as input data. The functionality of the tool seems to rely on a single type of input data (a small number of keypoints labeled using DeepLabCut) and throws away a large amount of information in the keypoint labeling step. If the main goal is to build a robust freezing detector then why not incorporate image data (particularly when the best set of key points does not include any limb markers)? Are the thresholds chosen for smoothing and convolution adjusted based on agreement to a user-defined behavior? Jitter is mentioned as a limiting factor in freezing classifier performance - does this affect human scoring as well? The use of a weighted average of body part velocities again throws away information - if one had a very high-quality video setup with more markers would optimal classification be done differently? What if the input instead consisted of 3D data, whether from multi-camera triangulation or other 3D pose estimation? Multi-animal data?

It is unclear where the manual annotation of behavior is used in the tool as currently stands. Is the validation module used to simply say that the freezing detector is as good as a human annotator? One might expect that algorithms which use optic flow or pixel-based metrics might be superior to a human annotator, is it possible to benchmark against one of these? For behaviors other than freezing, a tool to compare human labels seems useful. The procedure described for converging on a behavioral definition is interesting and an example of this in a behavior other than freezing, especially where users may disagree, would be informative. It appears that manual annotation doesn't actually happen in the GUI and a user must create this themselves - this seems unnecessarily complicated.

A major benefit of BehaviorDEPOT seems to be the ability to run experiments, but the ease of programming specific experiments is not readily apparent. The examples provided use different recording methods and networks for each experimental context as well as different presentations of data - it is not clear which analyses are done automatically in BehaviorDEPOT and which require customizing code or depend on the MiniCAM platform and hardware. For example - how does synchronization with neural or stimulus data occur? Overall it is difficult to judge how these examples would be implemented without some visual documentation.