On 2014 Dec 23, David Keller commented:

Innovative antiviral delivery systems for patients non-compliant with pre-exposure prophylaxis

Pre-exposure prophylaxis (PEP) is a reasonable option when adherence to condom use is imperfect. In order to improve compliance with PEP by reducing pill burden, antiviral medication could be incorporated into oral contraceptive pills prescribed to seronegative female partners in HIV-discordant couples. In cases where the discordant seronegative partner is male, other means of reducing the pill count required to deliver PEP could be used. For example, a seronegative older man with erectile dysfunction who engages in risky behaviors could be prescribed a phosphodiesterase inhibitor pill (such as sildenafil) which also includes the recommended antiviral PEP medications.

On 2015 Jul 24, Miguel Lopez-Lazaro commented:

I agree with Dr. Baker that the SMT does not explain important aspects of the disease, and that new models of carcinogenesis should be explored. The “stem cell division theory of cancer” is a new model of carcinogenesis that integrates some features from existing theories (SMT, TOFT, and CSC model) and may provide a better framework for understanding the disease. It can explain, for instance, the cellular origin of cancer, the age distribution of the disease, the existence of metastatic cancers in the absence of primary tumors, or why a variety of factors that do not directly cause DNA alterations (e.g., mechanical, physical, chemical or neural factors) can induce cancer or modify cancer risk. http://www.ncbi.nlm.nih.gov/pubmed/26090957<br> http://www.ncbi.nlm.nih.gov/pubmed/26097879

On 2015 Jan 23, Donald Forsdyke commented:

GC% DIFFERENCES NOT CAUSED BY CONVENTIONAL SELECTION

Its average base composition (GC%) is a characteristic of a biological species. The present work questions the view that GC% differences between species reflect responses to conventional selective pressures on organism function. Thus, the result “challenges the causes and possible functional roles (if any) of GC content variations in grass and Monocot genomes” (1). Likewise, in earlier work, the authors noted that “it is not clear why GC content in introns should also be selected for. Thus, we think that selective hypotheses are not clearly established and are currently insufficient to explain all the data adequately” (2).

To resolve this it would be interesting to examine the GC% values of sympatric, so-called “sibling species” (espèces jumelles, Geschwisterarten). Here phenotypic differences are minimal. Indeed, it has been shown that very small differences in GC% should suffice to spark speciation. These initiating GC% differences could later be obscured by pressures on the phenotype that affect GC%. But when such phenotypic differentiation was minimal, traces of these initiating events might remain (3).

Speciation is still mainly studied in complex organisms. Virus species that infect the same host cell have less scope for developing phenotypic differences and can be construed as sibling species. Indeed, we might recall that the mid-20th century revolution in molecular biology owed much to physicists who studied the simplest living systems – viruses that infect bacteria.

Considering the retroviruses HIV1 and HTLV1 that both infect CD4 T lymphocytes, we find that HIV1 has one of the lowest GC% values known and HTLV1 has one of the highest GC% values known. Large differences are also found when other related viral species share a host cell. Since minute GC% differences can initiate divergence into recombinationally isolated species, it can be assumed that, in the absence of major superceding phenotypic differences, these GC% differences have remained and expanded in HIV1 and HTLV1 (4).

Viewed from a selective perspective, an ancestral retrovirus by virtue of avoiding recombinational blending with its cell-mates, should have been able to develop sufficient functional variation “in sympatry” to achieve full speciation while retaining phenotypic characters needed for the shared intracellular environment. All this harkens back to Romanes who in 1886 proposed that initiation of divergence into species could precede subsequent phenotypic changes (5).

(1) Clément Y, Fustier M-A, Nabholz B, Glémin S. (2015) Genome Biology and Evolution. (In press) doi:10.1093/gbe/evu278

(2) Glémin S, Clément Y, David J, Ressayre A. (2014) GC content evolution in coding regions of angiosperm genomes: a unifying hypothesis. Trends in Genetics 30, 263-270.

(3) Forsdyke DR (2001) The Origin of Species Revisited. McGill-Queen’s University Press, Montreal.

(4) Forsdyke DR (2014) Implications of HIV RNA structure for recombination, speciation, and the neutralism-selectionism controversy. Microbes and Infection 16, 96-103 doi: 10.1016/j.micinf.2013.10.017.

(5) Romanes GJ (1886) Physiological selection: An additional suggestion on the origin of species. Journal of the Linnaean Society, Zoology 19, 337-411.

On 2015 Jan 29, GEANNCARLO LUGO commented:

Nice work. I truly enjoyed reading it. Looking forward to see the cover page!

On 2014 Dec 31, Walter Vogel commented:

For a detailed critique of this paper by @ClausWilke, see https://pubpeer.com/publications/33795E88EB9C62075F5194965C6874

On 2016 Oct 03, Duke RNA Biology Journal Club commented:

Overall, we were very impressed with the thoroughness of the paper - we rarely found an experiment that did not have orthogonal experimental validation along with the necessary controls. Leaving the discussion, we were convinced that NRAV plays a role in IAV host immune response and possibly a broader viral response, which control experiments hinted at.

More critically, we felt the paper focused on creating a broad understanding of NRAV’s function in immunity while leaving many loose ends in mechanistic understanding. For instance, an undeveloped part of this paper was the role of ZONAB in NRAV transcriptional repression. While the pulldown assays in Fig 6 I and J suggest ZONAB acts as a transcription factor of MxA, it’s not clear how this is regulated in the presence of NRAV. Perhaps performing a mass spec analysis of all proteins interacting with NRAV, see Fig 6H, would help uncover the rest of the mechanism. Additionally, our group had concerns with Fig 7. The determination of such a long RNA with structure prediction software is risky since longer RNA is more likely to be predicted to fold into many different structures with similar free energy compared to a much smaller RNA, the general query molecules for these programs. We would prefer the authors confirm these predictions with biochemical assays such as RNAse protection and SHAPE-based assays. We were also curious why the NRAV truncation mutants were not used to determine the ZONAB interaction site using pulldown methods.

To summarize, this paper scratches the surface of NRAV’s role in viral host immunity. We hope the authors continue to provide more depth to the story by determining how NRAV is regulated, how NRAV changes histone methylation patterns at ISG transcription sites and how ZONAB specifically fits into this response. We look forward to learning more about the role of lncRNA in the immune response and will be interested to see how the other lncRNAs from Fig 1A fit into this emerging field.

On 2015 Aug 19, Bill Cayley commented:

A good example of when "less" is "more": https://lessismoreebm.wordpress.com/2015/05/11/randomized-clinical-trial-of-the-effect-of-gum-chewing-on-postoperative-ileus-and-inflammation-in-colorectal-surgery/

On 2014 Dec 23, Torsten Seemann commented:

The PLACNET software can be found at http://sourceforge.net/projects/placnet/

On 2015 Jan 06, IGSJC - The International General Surgery Journal Club commented:

First author Adil Haider will be joining IGSJC to discuss "Incremental Cost of Emergency Versus Elective Surgery" on Twitter January 14-15, 2015. For the month of January, Annals of Surgery has opened access to the article (available at http://bit.ly/IGSJC_Jan15), and all are welcome to contribute to the discussion.

For tips on getting started, read http://igsjc.wordpress.com/guide-for-new-twitter-users/. Then, read the article, and return to Twitter on January 14 to take part!

Search on Twitter for #IGSJC (https://twitter.com/hashtag/igsjc) for the latest info and to join the chat.

On 2014 Dec 30, William Grant commented:

Here are some papers not included in this review showing benefits of vitamin D testing. Peiris AN, Bailey BA, Grant WB, Mascitelli L. Vitamin D testing. Lancet 2012 May 5;379:1699-1701.

Der T, Bailey BA, Youssef D, Manning T, Grant WB, Peiris AN.,Vitamin D and prostate cancer survival in veterans. Military Med. 2014;179 (1) :81–84.

Peiris AN, Bailey BA, Manning T. Relationship of vitamin D monitoring and status to bladder cancer survival in veterans. South Med J. 2013 Feb;106(2):126-30.

Bailey BA, Manning T, Peiris AN. Vitamin D testing patterns among six Veterans Medical Centers in the Southeastern United States: links with medical costs. Mil Med. 2012 Jan;177(1):70-6.

Of course some may have been published after your literature search was completed. However, they do support the benefits of vitamin D testing among hospital patients.

Also, this paper is supportive in that it found that patients in the ICU with very low 25OHD concentrations benefited from vitamin D supplementation. Amrein K, Schnedl C, Holl A, Riedl R, Christopher KB, Pachler C, Urbanic Purkart T, Waltensdorfer A, Münch A, Warnkross H, Stojakovic T, Bisping E, Toller W, Smolle KH, Berghold A, Pieber TR, Dobnig H. Effect of High-Dose Vitamin D3 on Hospital Length of Stay in Critically Ill Patients With Vitamin D Deficiency: The VITdAL-ICU Randomized Clinical Trial. JAMA. 2014 Oct 15;312(15):1520-30.

One of the impediments to acceptance of vitamin D seems to be the lack of supportive trials. The trials have not supported the ecological and observational studies largely because the trials have not been properly designed. Too often, people with normal to high 25OHD concentrations are enrolled and given a small amount of vitamin D. The proper way to conduct such trials was outlined recently in this paper: Heaney RP. Guidelines for optimizing design and analysis of clinical studies of nutrient effects. Nutr Rev. 2014 Jan;72(1):48-54.

Disclosure I receive funding from Bio-Tech Pharmacal (Fayetteville, AR) and Medi-Sun Engineering, LLC (Highland Park, IL).

On 2017 Feb 20, Stuart MacGregor commented:

Thank you for the bug report Christoph. Following your bug report (January 2016), the bug was fixed on the VEGAS2 web-based version in January 2016 but unfortunately not until January 2017 on the VEGAS2 offline version. Users who ran the top % test (not the default option) using VEGAS2 prior to these dates (or using the now retired VEGAS1) should re-run their analysis using VEGAS2. There is an FAQ item on this at https://vegas2.qimrberghofer.edu.au

On 2017 Jan 18, CHRISTOPH LANGE commented:

The VEGAS/VEGAS2 software can provide p-values for the top-percentage statistic that are anti-conservative, as the computation of the null-distribution is incorrect. In our technical report, we discuss the issue and provide code that, if included in VEGAS/VEGAS2, will provide correct p-values.

http://biorxiv.org/content/early/2017/01/17/101014

On 2015 Feb 10, Diederik W J Dippel commented:

Dear dr Radecki, in my view the best way to synthesize evidence is the Cochrane collaboration's approach. They have done this for IV thrombolytic treatment and will quite likely update the review of intra-arterial treatment. I would highly recommend reading that report which will undoubtedly offer more insight and detail. Diederik Dippel

On 2015 Feb 06, Ryan Radecki commented:

Post-publication commentary: "MR-CLEAN & the New Golden Age"

I, among many others, have been highly skeptical of thrombolytic therapy and its role in the treatment of acute ischemic stroke. As has been well-documented, a few trials were positive, many were neutral, and a few were stopped early for harm or futility. To most of us, this indicates a therapy for whom only a small subset of those treated are ideal candidates for benefit, and the margin between benefit and harm is razor thin.

In my previous posts, I’ve sighed wistfully at the hope of The Next Big Thing in stroke treatment – local endovascular therapy, akin to percutaneous coronary intervention. However, each major endovascular trial published in the New England Journal last year failed to demonstrate benefit.

MR-CLEAN is different. MR-CLEAN is rather unambiguously positive. To be zero or minimally disabled? The endovascular intervention is favored 12% to 6%. “Functionally independent”, a modified Rankin Scale of 0-2, favors endovascular intervention 33% to 19%. A number needed to treat of, apparently, ~8 for independence is nothing to scoff at.

But why?...

http://www.emlitofnote.com/2014/12/mr-clean-new-golden-age.html

On 2015 Feb 02, NephJC - Nephrology Journal Club commented:

This study was discussed on Jan 20th and 21st in the open online nephrology journal club, #NephJC, on twitter.

Introductory comments are available at the NephJC website. The discussion was quite detailed, with more than 45 participants, including nephrologists, fellows and residents with great insight provided by participation of the first author, Areef Ishani.

A transcript and a curated (i.e. Storified) version of the tweetchat are available at the NephJC website.

The highlights of the tweetchat were:

• The authors have undertaken a meticulous and well conducted epidemiological study to explain the risks of parathyroidectomy in the dialysis population, with research funding from the manufacturer of cinacalcet (which is an alternative to surgical parathyroidectomy).

• There were concerns raised about the lack of matched controls who did not undergo surgery (especially given the contrasting result from the previous study using USRDS data, generalizability of the post surgical hospitalization data and the lack of long term follow up. However, this study does represent a solid estimate of the immediate and 1-year mortality following parathroidectomy in this population. The discussion, and the comments from the author, also helped to explain the steps taken and the reasons behind these design decisions.

• The final question of the optimal management of secondary hyperparathyroidism still remains unanswered. Opinion was widely divided on the ideal study: most likely a randomized trial of surgical versus medical treatment of uncontrolled secondary hyperparathyroidism with clinically relevant outcomes, but with possible comparator arms being calcimimetic therapy, surgical parathyroidectomy or even placebo, suggesting a lack of consensus.

Interested individuals can track and join in the conversation by following @NephJC or #NephJC, or visit the webpage at NephJC.com.

On 2014 Dec 18, Tom Kindlon commented:

None

On 2014 Dec 18, Tom Kindlon commented:

The evidence is not there to recommend CBT and GET to improve employment outcomes in CFS

For over a decade now, some individual patients with Chronic Fatigue Syndrome (CFS)* in Great Britain and Ireland (and probably elsewhere) have been pressurised by insurance companies and occupational health professionals into undertaking graded exercise therapy (GET) and the form of cognitive behaviour therapy (CBT) that is based on scheduling increases in activity. This seems to have been largely due to hype around the efficacy of GET and CBT and extrapolations from subjective measures, as the evidence that such interventions are efficacious in restoring the ability to work is week.

Based on the information in Tables 1, 3 and the qualitative results from this paper, CBT and GET have again been recommended to occupational professionals in these workshops.

A lot of the evidence regarding CBT and GET and their effect on occupational outcomes in CFS has been summarised in a review (1). For some reason this is quoted sometimes as justifying claims it is evidence-based to say that GET and CBT have been shown to restore the ability to work in CFS. However the data is far less impressive. It is summarised in table 6 of that paper. The accompanying text says: "Among the 14 interventional trials with work or impairment results after intervention, there were too few of any single intervention with any specific impairment domain to allow any assessment of association."

The PACE Trial is by far the biggest trial of these therapies in the field. It shows neither CBT nor GET led to an improved rate of days of lost employment [Means (sds): APT: 148.6 (109.2); CBT: 151.0 (108.2); GET: 144.5 (109.4); SMC (alone): 141.7 (107.5)] (Table 2) (2). Neither CBT nor GET led to improvements in numbers receiving welfare benefits or other financial payments (Table 4). These results are in contrast to the self-reported improvements in fatigue, physical functioning and some other measures (3).

A major audit of Belgian CFS rehabilitation (CBT & GET) centres also gives real-world data on the issue (4). The sample size was large, with over 600 patients with a confirmed diagnosis of CFS (using the Fukuda et al. criteria (5)) taking part. It "comprised on average per patient 41 to 62 hours of rehabilitation" It found that "physical capacity did not change; employment status decreased at the end of the therapy." Again improvements were found in some self-reported measures.

It should be noted that a large assortment of abnormalities have been found in terms of the exercise response in CFS, with high rates of adverse reactions have been reported in patient surveys from CBT and GET, particularly with the latter, again putting in to question any recommendations of CBT and GET for CFS (6,7).

All in all, I question suggestions that occupational health professionals should be recommending CBT and GET to individuals with CFS.

• I'll use the term for consistency.

References:

(1) Ross SD, Estok RP, Frame D, Stone LR, Ludensky V, Levine CB. Disability and chronic fatigue syndrome: a focus on function. Arch Intern Med. 2004 May 24;164(10):1098-107. http://archinte.ama-assn.org/cgi/content/full/164/10/1098 or http://archinte.ama-assn.org/cgi/reprint/164/10/1098

(2) McCrone P, Sharpe M, Chalder T, Knapp M, Johnson AL, et al. (2012) Adaptive Pacing, Cognitive Behaviour Therapy, Graded Exercise, and Specialist Medical Care for Chronic Fatigue Syndrome: A Cost-Effectiveness Analysis. PLoS ONE 7(8): e40808. doi:10.1371/journal.pone.0040808

(3) White PD, Goldsmith KA, Johnson AL, Potts L, Walwyn R, et al. (2011) Comparison of adaptive pacing therapy, cognitive behaviour therapy, graded exercise therapy, and specialist medical care for chronic fatigue syndrome (PACE): a randomised trial. Lancet 377: 823-836.

(4) [Fatigue Syndrome: diagnosis, treatment and organisation of care] KCE Reports 88. (with summary in English). Accessed: 6th August, 2012. https://kce.fgov.be/publication/report/fatigue-syndrome-diagnosis-treatment- and-organisation-of-care

(5) Fukuda K, Straus SE, Hickie I, Sharpe MC, Dobbins JG, Komaroff A. The chronic fatigue syndrome: a comprehensive approach to its definition and study. International Chronic Fatigue Syndrome Study Group. Ann Intern Med. 1994 Dec 15;121(12):953-9.

(6) Twisk FN, Maes M. A review on cognitive behavorial therapy (CBT) and graded exercise therapy (GET) in myalgic encephalomyelitis (ME) / chronic fatigue syndrome (CFS): CBT/GET is not only ineffective and not evidence-based, but also potentially harmful for many patients with ME/CFS. Neuro Endocrinol Lett. 2009;30(3):284-99. Review.

(7) Kindlon T. Reporting of Harms Associated with Graded Exercise Therapy and Cognitive Behavioural Therapy in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome Bulletin of the IACFS/ME. 2011;19(2):59-111. http://www.iacfsme.org/BULLETINFALL2011/Fall2011KindlonHarmsPaperABSTRACT/ta bid/501/Default.aspx

On 2016 Jun 02, David Keller commented:

Information Theory versus FDA regulations governing colon cancer screening

The ColoGuard colorectal cancer screening test is composed of 7 different sub-tests, including an assay for occult fecal hemoglobin, 5 assays for specific DNA changes associated with malignancy, and an assay for normal DNA to standardize the sample. The results of these seven subtests are combined using complex equations to yield a Composite Score ("CS") ranging from 0 to 1000 and directly related to the likelihood of colon neoplasia. A CS of 183 was established as the threshold for further screening with colonoscopy; If the CS is larger than or equal to 183, the patient is sent for colonoscopy, and if the CS is lower than 183, the patient is spared further testing. [1,2]

The FDA prohibits the release of any of the test data generated by the ColoGuard system except the single bit of information in the final result: positive or negative, over or under the threshold, yes or no to colonoscopy, 1 or 0. This one bit of information represents the distillation of the combined results from the seven sub-tests, but much of the information in those test results is lost in the process of determining the value of that single bit.

Now, consider the composite score, which can vary from 0 to 1000. How many bits of information does the CS convey? The answer is nearly 10 bits of information, because 10 bits can represent 0 through 1023 in decimal numeral, or 000000000 through 111111111 in binary. So, we start with a composite score which requires 10 bits of information to quantify the patient's risk of colon cancer, and we finish by distilling the result down to a single bit of information: positive or negative. In so doing, we have thrown away 9 perfectly good bits of information.

A negative ColoGuard result conveys just one bit of information about the status of the patient's colon epithelium, informing us only that the composite score was in the range of 0 to 182, a range which requires almost 8 more bits to specify precisely. The FDA forces Exact Sciences to throw away almost 7 perfectly good bits of information about the patient's colon epithelium. These discarded 7 bits tell the patient whether his composite score was 0 or 182 or somewhere in between, given an overall negative screening result.

The purpose of the FDA regulation is to force clinicians to adhere strictly to the decision threshold for colonoscopy of 183, which has been validated in a large clinical trial. But what of the patient who develops "soft" or borderline indications for colonoscopy some time later. The clinician may be undecided whether the patient's borderline signs and symptoms warrant a diagnostic colonoscopy or not. Knowing that the patient's recent composite score had been 17 or 170 could help tip the decision one way or the other.

We may learn over time that composite scores can exhibit informative trends over time. What is the significance of a composite score rising from 15 to 160 over 3 years? We will never know if we throw away the 7 bits that let us quantify the composite score with precision.

We have discussed the loss of clinical information which occurs when the ColoGuard composite score is distilled down to one bit. There is also substantial loss of information when the Cologuard sub-test results are crushed together to derive the composite score. In my next comment, I will discuss some effects of the loss of information which occurs related to ColoGuard's fecal hemoglobin concentration sub-test. [3] We will learn how the FDA's prohibition of patient access to Cologuard's internal sub-test results could be harmful to patients.

References

1: Imperiale TF, Ransohoff DF, Itzkowitz SH, Turnbull BA, Ross ME; Colorectal Cancer Study Group. Fecal DNA versus fecal occult blood for colorectal-cancer screening in an average-risk population. N Engl J Med. 2004 Dec 23;351(26):2704-14. PubMed PMID: 15616205.

2: Imperiale TF et al. Online Supplement to the above print article, accessed on 4/18/2016. http://www.nejm.org/doi/suppl/10.1056/NEJMoa1311194/suppl_file/nejmoa1311194_appendix.pdf

3: PubMed Commons comment: http://www.ncbi.nlm.nih.gov/pubmed/27235008#cm27235008_15979

On 2016 Mar 13, Tamir Tuller commented:

We (and everyone) know that correlation on binning may be misleading if reported in a non-transparent way; specifically binning tends to increase the correlation (but usually has a much weaker effect on the p-value). However, we frankly do not understand why this ‘lecturing’ about the topic appears here (next to our study). Our study includes various sophisticated statistical tests, the binning procedure is described at the beginning in a coherent manner, the correlation for various bin sizes is also reported (one can learn about the relation between binning and correlation simply by looking at figure S5 without the need for this unnecessary correspondence :-)), etc. Thus, we believe that the nature of the signal and the challenging data should be very clear to readers who thoroughly read the paper (but we guess that it may be misleading, as any other paper would, if you do not bother to read all the details :-)).

The statistical analysis in papers in our field (if they are performed accurately) should consider various aspects including non-trivial biases in the data, discretization, various confounding variables/explanations, various aspects of molecular evolution, huge datasets, etc. Thus, the reader and not only the author should consider them when evaluating the results; specifically, the strength of a correlation should be evaluated in the light of all these aspects. The aim of mentioning other papers and ‘top statisticians’ was to demonstrate that there are many people (as opposed to Plotkin/Shah/Cherry) that do understand this point.

If the number of points in a typical systems biology study is ~300, the number of points analyzed in our study is 1,230,000-fold higher (!); a priori, a researcher with some minimal experience in the field should not expect to see similar levels of correlations in the two cases. Everyone also knows that increasing the number of points, specifically when dealing with non trivial NGS data, also tends to very significantly decrease the correlation. The aim of the binning was to align our signal to previously reported signals in the field (in terms of number of points), and as mentioned the paper includes many other analyses that give the reader a greater context for the signal (including an explicit graph reporting the relation between bin size and the correlation); in addition the non-binned correlation (0.02-0.07) is comparable to the level of correlation between two Hi-C measurements (~0.05) from different labs (!). It is clear that a typical signal in our field (e.g. higher than the correlation of 0.12 or even the “high” 0.38 mentioned in your paper Weinberg DE, 2016) if transferred via such a noisy/biased ‘channel’ with increased number of points will be order of magnitudes lower than our non-binned data.

We, of course, do not expect that further back-and-forth will convince Plotkin and Shah of our points. But hopefully this exchange will at least have some value to the field for scientists who work to draw inferences from genomic datasets; specifically, we hope that other scientists will learn to thoroughly consider all the aspects mentioned above and below when reading/writing a scientific paper.

BTW: regarding the correlation 0.12 that was improved to 0.38 in the new study. In the new study (Weinberg DE, 2016) you still did not perform many of the required statistical controls (among others control for Kozak sequence and AA bias) according to our review [http://www.cs.tau.ac.il/~tamirtul/Shah_et_al_review.pdf].

Tamir Tuller & Alon Diament, Tel-Aviv University, March 13, 2016

On 2016 Mar 09, Joshua Plotkin commented:

Hopefully this exchange now illustrates that correlations on binned data are terribly misleading. In the paper by Diament et al 2014, the authors never reported the actual correlation (r = 0.022) between two genomic measurements; instead they reported correlations on binned data (r = 0.86). As a result, the amount of variation in 3D position explained by codon usage, which is the central claim of the study, was inflated by 1,500-fold.

There is no need to consult “top professional statisticians” to understand the obvious fact that binning data tends to inflate correlations, and that it has no scientific justification regardless of the size of a dataset.

Diament and Tuller offer one non-scientific justification in their reply: that previous publications in “top journals” have done the same thing, citing Ghaemmaghami S, 2003 and Shah P, 2013 as examples. Even if this were true, appealing to journal name is not a strong justification for repeating statistical errors. Moreover, in fact, both cited studies used binning only to graph the data, showing also the variation within each bin, whereas the correlations were calculated on the unbinned data.

We agree with Diament and Tuller that r = 0.12 should not have been described as a “strong” correlation by Shah P, 2013. (The same analysis on an improved experimental dataset yields r = 0.38, Weinberg DE, 2016). However, unlike Diament and Tuller, we maintain that scientists must nonetheless report the actual correlations between measured quantities, instead of inflating correlations by binning, which would have produced a misleading r = 0.62 in the case of Shah et al. 2013.

We do not expect that further back-and-forth will convince Diament and Tuller of these points. But hopefully this exchange will have some value to the field of scientists who work to draw inferences from genomic datasets.

--Joshua Plotkin & Premal Shah, University of Pennsylvania, March 9 2016

On 2016 Mar 05, Tamir Tuller commented:

Plotkin couldn’t (and shouldn’t) be a reviewer of our paper simply since he was in our list of reviewers to exclude when the paper was submitted (unfortunately, indeed Plotkin’s & Shah’s comment here demonstrates that many of the details in our paper were completely overlooked by them in the biased ‘review’ they provide here, as all the raised issues were thoroughly answered in the original manuscript). Specifically in our data there were 369,000,000 initial points that were binned to up to 64,000 (!) bins (not 2 points! :-) ). The comment and answer to Cherry actually include all the important details regarding Plotkin’s & Shah’s comment. As noted, the other reviewers (and bio-statisticians that were presented the study) were, of course, aware of the binning process and found the paper interesting, correct and worthy of publication. We think that a discussion about binning (that it seems that Plotkin & Shah would like to promote) should actually include their own recent study (Shah P, 2013): there you can learn among others that in Shah P, 2013 Plotkin & Shah report in the abstract a correlation which is in fact very weak (according to their definitions here), r = 0.12, without controlling for relevant additional fundamental variables, and include a figure of binned values related to this correlation. This correlation (0.12) is reported in their study as “a strong positive correlation”. It is also important to mention that the number of points used for computing this correlation is more than one order of magnitude lower than the number of points/bins related to some of the correlations we report.

Given their comment here, it is probable that Poltkin & Shah’s paper would not have been published (at least in its current form) had they reviewed it themselves. :-)

Our full critical comment on (Shah et al., 2013) can be found here: Shah P, 2013 (or here http://www.cs.tau.ac.il/~tamirtul/Shah_et_al_review.pdf ).

A point by point answer to Plotkin’s & Shah’s ‘review’ can be found here http://www.cs.tau.ac.il/~tamirtul/Plotkin__reply.pdf

Our full answer to Cherry’s, Plotkin’s & Shah’s claims regarding binning can be found here: http://www.cs.tau.ac.il/~tamirtul/Cherry_reply.pdf

Tamir Tuller & Alon Diament, Tel-Aviv University, March 5, 2016

On 2016 Feb 04, Joshua Plotkin commented:

Weak effects made to appear strong by inflated correlation coefficients (reply)

The titular claim of this paper is that codon usage and gene function are “strongly correlated” with physical proximity within the cell. To support this claim, the authors report correlations between various features of genes, such as codon usage, expression level etc, with 3D genomic organization data. Unfortunately, as Joshua Cherry has pointed out on PubMed commons (Nov 24, 2015), these correlations are all based on binned datasets. By systematically binning the data, the authors remove noise and artificially inflate the strength of the correlation, so that the resulting correlation coefficient does not reflect the actual strength of correlation in the data. In the extreme case of n=2 bins, for example, all the correlations would be r=1.

We raised these concerns — the exact same ones Joshua Cherry expressed post-publication — in our original review of the submitted manuscript. We submitted our review to Nature Communications on May 21, 2014, and we never heard back from the journal or saw a revised version of the manuscript. It is now clear that our concerns were ignored both by the authors and by editors at Nature Communications.

Artificially inflating correlations by binning data is a serious issue in ongoing biological studies. By posting our original review of this manuscript, alongside Cherry's post-publication critique, we hope to draw awareness and open discussion regarding this scientific issue.

Joshua Plotkin & Premal Shah, University of Pennsylvania, February 4, 2016

----ORIGINAL REFEREE REPORT SUBMITTED May 21 2014----

Remarks to the Author:

The manuscript by Diament et al. aims to understand how the three dimensional arrangement of a eukaryotic genome is organized. The natural hypothesis is that functionally related genes are positionally closer to each other in space -- a hypothesis that has been proposed earlier but with limited empirical support. Here, the authors claim that earlier studies failed to identify a strong relationship between position and function due to lack of appropriate metrics to assess functional similarity between genes. The authors propose a "novel" metric based on patterns of codon usage and they demonstrate a putatively strong relationship between functionally related genes and their proximity in 3-D.

The manuscript is severely flawed from both biological and statistical stand-points, and is not fit for publication. Following are my detailed comments:

1) The "novel" CUBS method proposed by the authors is not novel at all. There is a rich literature of using Kullback-Liebler based distance metrics for studying patterns of codon usage, which the authors have completely ignored. Although, their metric is a symmetric version of KL-distance, the entire basis of this metric is not novel.

2) The entire analysis is predicated on the assumption that functionally similar genes have similar patterns of codon usage. The only work cited in support of this notion is De Bivort et al 2009, where those authors found that amino acid metabolism might play a role in affecting protein composition for certain functionally related proteins in yeast. However, even there the extent of this effect was found to be limited to 20,000-60,000 residues, which constitutes less that 2% of the entire genome. To say that this pattern holds not only across the entire genome but also across four other species demands quite a stretch of imagination.

3) More importantly, every single correlation reported in the paper is based on binned data. Although it is sometimes appropriate to bin the data for visualization purposes, it is entirely without merit to report correlation coefficients (and associated p-values) on binned data. This fact id demonstrated by comparing figures 3D and S2A, where changing the bin-size effects the apparent "correlation". All the correlation calculated should be based on the raw data and "improving statistical accuracy" is not a vlid justification for arbitrarily binning data. This problem of correlation-inflation due to binning the data is quite serious (eg Kenny & Montanari J Comput Aided Mol Des. 2013). Based on their own figures 3D and S2A, it seems clear that their results either have very small effect or do not at hold at all when analyzing the actual raw data.

4) This statistical problem (#3) is further compounded by the fact that each data point in the correlations is based on a pair of genes, and hence the points are not independent of each other -- whereas t-tests used to assess significance assume independence. The standard way to deal with such data is to use Mantel's test or one of its several derivatives. The authors also need to take into account issues related to multiple regressions. Its likely that several of the gene features used as proxies for function are, again, highly correlated with each other. Once again, the nominal statistical significance of the results is inflated by failing to account for these dependencies.

5) Finally, the authors make no attempt to explain why the data in their plots are so non-linear and even non-monotonic? Its clear that in several of the plots the relationship between CUBS and the predictor is highly non-linear, and that the linear fit is extremely poor. However, the authors make no effort to explain or even understand these patterns. Could they be an artifact of the data? If so, how does it affect the results?

6) Moreover, the correlation coefficients reported in most of their plots make no sense whatsoever. For instance, in Fig1B, the best-fit regression line of CUBS vs PPI barely passes through the bulk of the data, and yet the authors report a perfect correlation of R=1.

The issues above are so severe (starting from the unjustified assumption that similar codon usage implies similar gene function, which underlies the entire study) that revision cannot possible rectify these problems.

Remarks to the Editor: none

On date unavailable, commented:

None

On 2016 Mar 08, Joshua L Cherry commented:

In their response to my comment, Diament and Tuller not only attempt to defend their use of a procedure that dramatically inflates correlation coefficients, but advocate for its wider use, which would be most unfortunate. Nothing in their response justifies this procedure or refutes the warnings against it that I cited. I address their points briefly below, and in more detail here.

Statistical Significance vs. Effect Size

The response makes much of the high statistical significance of the correlations. This does nothing to address my comment, which was about effect size. A weak correlation, however statistically significant, is not a strong correlation.

Large Sample Size

The response emphasizes the large number of data points in the authors’ data set, but this in no way justifies their correlation-inflating procedure. The large size of the dataset only allowed the authors to bin and average large numbers of data points, which only exacerbated the problem by leading to a more dramatic inflation of the correlation coefficients. The response also seems to imply that a large sample size will cause a strong correlation to yield a low correlation coefficient, which is incorrect.

Other Evidence

The response claims that "The conclusions of our paper have also been tested in a recent study (Diament A, 2015), where we showed that 3D distances predicted by CUFS can be employed to reconstruct an improved 3D model of the yeast genome." If CUFS data did improve estimates of 3D distance, this would not justify or vindicate the inflation of correlation coefficients by binning and averaging. Furthermore, as explained in my more detailed response, there is little evidence that the 3D distance estimates were actually improved.

Measurement Errors

The response repeats the article’s argument about measurement noise, but this argument is flawed. It is true that a strong underlying correlation coupled with sufficient measurement error could produce a weak apparent correlation with the observed properties. It does not follow that a weak correlation with these properties implies a strong underlying correlation. Counterexamples are common--probably the rule rather than the exception--and include Francis Galton’s classic studies of height among relatives. Galton and Pearson could easily have arrived at larger “correlation coefficients” through binning, but doing so would have defeated their purpose and been a great setback for statistics and quantitative genetics.

Diament and Tuller claim, based on the low correlation coefficient between two sets of yeast 3D distances (r=0.05), that the maximum possible CUFS/3DGD correlation is 0.05. Were this correct, it would only argue for an underlying correlation of 0.022/0.05=0.44 for yeast, far short of the reported r=0.86. The claimed maximum is in fact incorrect, for several reasons:

• If the correlation coefficient between replicate datasets were 0.05, the correct maximum would be sqrt(0.05)=0.22 (the inferred correlation of either replicate with reality), from which one can argue only for an underlying r=0.1.

• Diament and Tuller did not analyze data from replicates, but compared the data that they used in the article to values derived from an older experiment based on a different technique. The low correlation between the datasets may be due mainly to noise in the other dataset.

• The original response submitted by Diament and Tuller included two other comparisons that did not suffer from the above problem, and these yielded much higher estimates of reliability: r=0.39 and r=0.54. This information was omitted from the version of the response posted by Tuller.

Diament and Tuller argue that r=0.05 for the two datasets, calculated as Spearman intended, absurdly implies that "any attempt to study 3D genomic organization using Hi-C...is futile”. In reality it implies only that these particular datasets are not very similar, which is not absurd, and undoubtedly true. Their apparently preferred alternative--binning the data to obtain a high correlation coefficient and concluding that the measurements are very similar--is clearly incorrect.

Concluding Remarks

The correlations between 3DGD and CUFS, though statistically significant, are quite weak, and should have been reported as such. The available information does not support the contention that they reflect strong correlations made weak by measurement noise. That others have committed or overlooked a "common error" is not an argument in its favor. That correlations coefficients in a field are often weak is no reason to inflate them or describe weak correlations as strong. It would be one thing to argue that a correlation coefficient of 0.022 tells us something important, but it is quite another to report an inflated correlation coefficient of 0.86 and describe the correlation as strong, as done by the authors.

On 2016 Feb 27, Tamir Tuller commented:

We are happy to report that the paper, including the binning procedure, was presented to top professional statisticians who found the paper very interesting. The comment by Cherry was submitted to Nature Communications as a comment, and was reviewed by at least two professional reviewers. Based on the review the decision was that the comment is not required as it does not teach us anything new and is irrelevant/incorrect in this context. Specifically, it is known that binning tends to increase the correlation; however, the opposite is also true, when there are more points the correlations tend to be lower. In the analyzed data, the binning procedure was described very clearly and mentioned at the beginning of the paper; there were 369,000,000 initial points that were binned to up to 64,000 (!) bins (not 2 points!) so that the correlations will be comparable to correlations reported in all previous systems biology studies in recent years (actually the number of points is still orders of magnitudes (!) higher than previous systems biology studies that we know of). As described in the link to the detailed reply (see below) the obtained correlations with raw data are similar to the ones obtains between two Hi-C measurements, demonstrating that indeed given the nature of the data the correlations are very high. A detailed reply to Cherry’s comment can be found here: http://www.cs.tau.ac.il/~tamirtul/Cherry_reply.pdf. In this link, we further explain why Cherry’s claims were thoroughly addressed in the original manuscript, and that the methods and results were presented in a transparent manner. Most importantly, we reiterate that the relation between variables has been subject to stringent statistical tests, and that the observed signals are indeed strong with respect to expected and previously reported ones in large-scale genomic studies. In addition, we illustrate again that the reported correlations are comparable to the maximal correlation expected when comparing large scale noisy data after quantization. Finally, we show that the correlations reported in our study are similar to the correlations obtained between two Hi-C experiments; thus, if we follow Cherry’s line of thought, we actually should absurdly conclude that the Hi-C protocol in general is problematic. We discuss the generality of our conclusion to systems biology analysis of Next Generation Sequencing (NGS) data.

On 2015 Nov 24, Joshua L Cherry commented:

Weak effects made to appear strong by inflated correlation coefficients

Diament et al. claim that the three-dimensional distances between eukaryotic genes are strongly correlated with differences in codon usage and gene function. These claims are based on correlation coefficients calculated by an illegitimate procedure that grossly inflates their magnitude. The correlations are in reality very weak, and the article’s conclusions are therefore unjustified.

Diament et al. report high values (0.74-0.96) for Spearman’s rank-order correlation between measures of codon usage dissimilarity (CUFS) and 3D distance between genes (3DGD). These impressive values, however, are not the true correlation coefficients between the variables. Rather, they are the correlation coefficients of average values for bins of thousands of data points with similar values of CUFS. This binning and averaging procedure can be expected to yield high values in the presence of very weak correlations. The authors describe the process as "reducing biological noise through averaging". In reality it suppresses much of the variation in one variable that is not explained (in the statistical sense) by the other, making it appear as though the explained variation is a larger fraction of the total, and inflating the correlation coefficient accordingly. A miniscule correlation can be made to look like a strong correlation with this procedure, so long as the expected value of one variable mainly increases with the value of the other. Computing correlations based on averages is described by one textbook [1] as a “common error” that “can easily lead to an inflated correlation coefficient”, and correlations inflated in this way have been criticized elsewhere [2,3].

Supplementary Fig. 5 of the article suggests that the true correlations between CUFS and 3DGD, corresponding to one data point per bin, are quite weak. Using data provided by the authors, I have found that they are very weak indeed: the Spearman’s correlation coefficients are 0.019, 0.022, 0.025, 0.071, and 0.034 for S. pombe, S. cerevisiae, A. thaliana, M. musculus, and H. sapiens respectively. The central claims of the article are therefore unfounded, as the reported strong correlations are an artifact of the binning procedure.

The weakness of the correlations is evident in the meager sensitivity of averaged 3DGD to CUFS, which is apparent in Fig. 2 of the article. For example, for S. cerevisiae, for which a correlation coefficient of 0.85 was reported, averaged values of 3DGD vary only from ~3.0 for the lowest CUFS values to ~3.1 for the highest (a few outliers approach 3.3, but, as the authors argue, these are not meaningful). This is only a small fraction of the variation in 3DGD values, which range from 1 to 13 with a standard deviation of 0.72.

Another perspective is provided by the distributions of CUFS values for different 3D distances. Distributions in S. cerevisiae are shown in here for distances between 1 and 5, which encompass 99.8% of the gene pairs (for larger distances the distributions vary more widely but are dominated by pairs involving one or a few genes). These distributions are quite similar to one another. Indeed, for 3DGD < 5 (encompassing 98% of the gene pairs), the distributions are difficult to distinguish, and the only distinguishing feature for 3DGD = 5 is due to effects of just a few genes (see figure legend). Differences between the means of the distributions (shown graphically in the plot) are quite small compared to the variation of CUFS within each 3D distance category. Clearly CUFS is not strongly associated with 3D distance.

The weakness of the correlations is in no way negated by the fact that the authors’ binning procedure yields high values. Most correlation coefficients will be inflated by this procedure, including the prototypical coefficients calculated by Galton and Pearson. For samples drawn from a bivariate normal distribution with a correlation coefficient of just 0.03, this procedure, with the relevant bin and sample sizes, yields Spearman’s correlation coefficients greater than 0.9. It would be absurd to describe such variables as strongly correlated, and reporting a Spearman’s correlation of >0.9 would be grossly misleading.

Large data sets can give us the statistical power to detect very weak correlations, revealing what has been called the “crud factor”: that “everything correlates to some extent with everything else.”[4] Weak correlations are easily produced in the absence of a direct or otherwise interesting connection between the variables. Weak correlations are sometimes enlightening, but exaggerating their strength only obscures.

References

1. Triola, M. F. Elementary Statistics. Addison-Wesley, Reading, MA (1992)

2. Kenny, P. W. & Montanari C. A. Inflation of correlation in the pursuit of drug-likeness. J. Comput. Aided Mol. Des. 27:1-13 (2013).

3. Brand, A. & Bradley M. T. More voodoo correlations: when average-based measures inflate correlations. The Journal of General Psychology. 139(4):260-272 (2012).

4. Meehl, P. E. Why summaries of research on psychological theories are often uninterpretable. Psychological Reports. 66:195-244 (1990).

On 2015 Jan 22, Yuriy Pankratov commented:

This discussion could be more enlightening and even reach a consensus of some sort if our respected opponents, instead of making ungrounded accusations and avoiding inconvenient facts, tried to address the most serious issues, raised in our comments at PubMed Common and JNS website. These issues include: stark contradiction between the EGFP and LacZ expression phenotypes shown in Fujita et al. and data shown in previous publications (all of which went through rigorous peer review by the way); lack of direct evidence of notable impairment of synaptic transmission in dnSNARE mice and existence of clear evidence of the opposite; large pool of evidence supporting physiological role of astroglial exocytosis which does not rely on the dnSNARE mice at all. Neither paper itself nor Authors’ responses to comments (which basically repeats what was said in the article) address these issues.

Still, we think that some consensus might be found. Before going to that point, we would like to clarify points raised by our opponents in their last post. 1) For the sake of unbiased discussion, citing one paper showing a lack of VAMP2 expression in astrocytes (Schubert et al.2011) one might mention at least one paper from the large pool showing the opposite (Martineau et al 2013).

More importantly, one should not swap between quantitative and “all-or-none” kind of reasoning to one’s convenience. If we assume that level of astrocytic expression of VAMP2 of tenth of that in neurons is low enough to make VAMP2 non important for function of astrocyte, than we have to assume the existence of certain level of expression below which dnSNARE transgene will not significantly affect neuronal function as compared to astrocytes. OK, it may be not tenth but hundredth fraction, dose not matter.

We thankful to our opponents for bringing up an example of tetanus and botulinum toxins. Even theses deadliest toxins act in dose-dependent manner. Both on the levels of whole organism and single presynaptic terminals, smaller doses of these toxins (as compared to LD50 and IC50) have milder effects. So, it is very likely that effects of dnSNARE expression are dose-dependent (if not to believe in homeopathy, of course).

The same is applicable to the action of doxycycline, which is also dose-dependent. So one could not expect 100% inhibition of transgenes, especially at oral administration of Dox. To answer first part of opponents comment 2), the Figure 1O-R from Halassa et al. shows efficient, but incomplete suppression by Dox, rather than “leaky” EGFP expression. To what extent the same is applicable to Fig.2C of Fujita et al, let the reader to decide. Of course, non-complete suppression by Dox is a downside of tetO/tetA system but this can be easily remedied by comparing On-Dox and Off-Dox data.

2) Theoretically speaking, concern that “neurons express the dnSNARE transgene at all “ may be applicable to any glia-specific transgenic mice. One could not a priori expect an absolute specificity of expression of neuronal and glial genes, the data of Cahoy et al. 2008 are the good illustration. This, rather philosophical, question goes far beyond the current discussion. There is no molecular genetic tool to ensure 100% glial specificity. On practice, one could only expect to obtain a negligible (again, in relative sense) level of neuronal transgene expression and verify the lack of significant impact on neuronal function.

3) Regarding the putative “dramatic and unpredictable “ effects of neuronal dnSNARE expression, the TeNTx and BoNT give a good indication of what to expect. However, dnSNARE mice do not show any notable deficit of motor or respiratory function. On a level of synapses, there was no evidence of any significant decrease (not saying about complete inhibition) of vesicular release of main neurotransmitters (Pascual et al. 2005; Lalo et al. 2014). On contrary, our data show an impairment of signals triggered by activation of Ca2+-signalling selectively in astrocytes (Lalo et al. 2014; Rasooli-Nejad et al. 2014). Let it to the reader to decide, to what extent available functional data support the opponents’ notion that “synaptic transmission may directly be suppressed by dnSNARE expression in neurons “ and that “Even very low levels of expression of dnSNARE in neurons invalidate any conclusion based on this transgenic mouse “.

One might argue that dnSNARE transgene could be expressed only in the certain subset of neurons or in some specific brain region thus strongly affecting some specific function rather than causing general, milder, functional deficit. However, this is unlikely for the supposed basal leakiness of the tet-off system and further experiments would be required to identify such regions/neuronal subsets.

4) Addressing the second half of the point 2) – One can only wonder why, in 2012, already knowing that their results contradict to data presented by that time by several studies, our respected opponents did not contact authors of those publications to request mice from them? Again, one might only wonder why PCR data generated from 2 batches of mice have sample size of n = 3 – 4 (meaning 1-2 tissues per batch) ?

5) Regarding the intrinsic limitations of dnSNARE mice, anyone working with them is aware of fact that EGFP, LacZ, and dnSNARE genes were inserted independently. However, their expression is controlled by the same factors so their expression probabilities depend on the same set of parameters and therefore are not truly independent, from mathematical point of view. The correlation in expression of these transgenes is supported by the co-inheritance. Furthermore, data of Halassa et al. show that 97% of cells expressing the dnSNARE, also express EGFP. We would like to emphasize that the opposite - the presence of true mosaic expression pattern in dnSNARE mice, i.e. existence of number of individual cells expressing dnSNARE and not expressing EGFP and number of EGFP-only cells, has not be shown so far; Fig.3 from Fujita et. al 2014 does not show this either.

Thus, even assuming the leakiness of the “tet-off” system, one might expect probability of EGFP expression to be of the same order of magnitude as that of dnSNARE, this is also agrees with data of Fujita et al. So, in case of absence of EGFP expression in a large population of neurons, the presence of even small fraction of neurons expressing dnSNARE is very unlikely. From mathematical point of view, the probability of certain population of neurons to express only dnSNARE will fall exponentially with the expected size of population.

Finally, one could hardly deny the large difference in the phenotype of the cohort of dnSNARE mice, described by Fujita et al. and the cohort of mice used by other groups. The point of some consensus could be that in some, still unidentified conditions, the tetA/tetO system may suddenly became leaky, causing some level of expression GFAP-driven transgenes dnSNARE, EGFP and lacZ genes in neurons. So, in experiments with dnSNARE mice extra care should be done to verify the lack of neuronal dnSNARE expression. This can be done by showing the absence of surrogate reporters EGFP or lacZ in neuronal populations of interest combined with electrophysiological data showing the lack of deficit of synaptic neurotransmitter release. This could be a good practice for any glia-specific inducible transgene.

On 2015 Jan 13, Maiken Nedergaard commented:

With all respect to the commentator, the misleading posted arguments and evident lack of insight into the biology and construction of the dnSNARE mice underscores the danger of such unsolicited and unreviewed posting, not subject to peer review. That said, we have clarified several points below, and will leave it to the reader to evaluate our data as published in Fujita et al., J Neurosci, 2014 and to form an unbiased opinion.

1) Our immunolabeling data does not show that astrocytes fail to express VAMP2, nor did we ever make such a claim. Rather, our analysis documents that astrocytes express VAMP2 at much lower levels than neurons, both in vitro and in vivo. Besides our own data, this notion is supported by independent sets of data – transcriptional analysis showed that astrocytic expression of Vamp2 is a tenth that of neurons (Cahoy JD, 2008), while an immunohistochemical analysis showing that VAMP2 is enriched in presynaptic terminals, and is not detectable in astrocytes (Schubert V, 2011).

The dnSNARE mice express the 3 genes (EGFP, LacZ, and dnSNARE) independently of one another; their expression cassettes were injected in the oocyte as separate genes. Thus, a mosaic pattern of expression is to be expected, and that is what Fig. 3 documents. We certainly agree that astrocytes express all 3 transgenes at higher levels than neurons do; we documented precisely this point in Fig. 2D. However, the key element of our study is to demonstrate that leaky neuronal expression, and hence low levels of dnSNARE expression in neurons, may significantly impact synaptic transmission, since VAMP2 is essential for the fusion of synaptic vesicles with the presynaptic membrane.

Incidentally, the best example of the damage that may be wrought by any neuronal expression of dnSNARE is the acute mortality associated with nanogram quantities of botulinum toxin. Botulinum toxin’s actions are analogous to those of dnSNARE; the functional consequences of both derive from the necessity of the SNARE complex to the release of synaptic vesicles.

2) We will leave it to the reader to evaluate the immunohistochemical analysis, but will note that the leaky expression of EGFP was also displayed in Halassa MM, 2009, Fig. 1O-R. Unfortunately, EGFP expression has no predictive value for expression of the dnSNARE transcript, since as noted, the genes are independently expressed. We received the dnSNARE mice from Dr. McCarthy, rather than one of the groups using the mice, because Dr. McCarthy made the dnSNARE mice. We have analyzed two independent shipments (received in 2007 and 2012) and identified neuronal expression of EGFP in both sets of mice.

3) This comment reveals a basic misunderstanding of the inherent limitation of the dnSNARE mice. Direct inhibition of vesicular fusion in neurons is expected to interfere - in dramatic and unpredictable ways - with neuronal activity in the intact animals. It is not a matter of the relative expression of dnSNARE expression in neurons versus astrocytes. Even very low levels of expression of dnSNARE in neurons invalidate any conclusion based on this transgenic mouse. Moreover, we document in Fig. 2H that neurons express the dnSNARE transgene in the presence of doxycycline, and that transgene expression is increased after removal of doxycycline. In other words, the expression of the dnSNARE transgene is also controlled by doxycycline in neurons.

4) Again, it is not a matter of the relative leakiness of dnSNARE expression in neurons. The fact that neurons express the dnSNARE transgene at all is the fundamental concern. Interfering with synaptic release is one of the most powerful manipulations that may be executed upon a neuronal network.

5) Once again, the point of our study is not to show predominant neuronal dnSNARE expression. We document that cortical neurons do express the dnSNARE transgene, and that the expression of dnSNARE in neurons is regulated by doxycycline. These observations invalidate the use of dnSNARE transgenic mice in the study of neuroglia signaling: It is not possible to attribute change in neuronal activity to gliotransmitter release, since synaptic transmission may directly be suppressed by dnSNARE expression in neurons.

6) Unfortunately, exchange speaks as much to the deficiencies of the PubMed Comment mechanism as to the self-defensive but ultimately unsupportable arguments of the commentator, as it invites a Reddit-like forum, in which informed and uniformed commentary are admixed into an unenlightening whole.

Commented by Maiken Nedergaard, Takumi Fujita, Michael J. Chen

On 2015 Jan 05, Yuriy Pankratov commented:

Article by Fujita et al. aims to put under question an importance of SNARE-dependent vesicular gliotransmission basing on their observation of “widespread “ neuronal expression of dnSNARE transgene. However, there are several inconsistences and misinterpretations in the data which seriously undermines the main conclusion on the paper.

1) The immunocitochemical analysis (Figure 3A-D), which is supposed to show “widespread neuronal expression” of b-Gal and EGFP in dnSNARE mice actually gives merely an anecdotal evidence of their expression in minor population of neurons. Neither Figure 3 nor the description of results in the text provides any statistical data. In Fig.3A, white arrows are used to draw attention to the correlation of green VAMP2- and orange MAP2-staining in neurons which is a trivial result. However, careful inspection of these images reveals a number of green puncta not associated with orange staining which might be as well located in astrocytes. To rule this out, the quantitative data on correlation of VAMP2 with neuronal and glial markers are essential. The lack of VAMP2 staining in cultured astrocytes (Fig3) is in controversy with data of Pascual et al (2005). This fast was ignored by the Authors, as well as other evidences of VAMP2 expression in astrocytes (Parpura 1995, Maienschein 1999, Montana 2004, Martineau 2013).

Similarly, careful inspection of images shown in Fig 3C,D shows that most of cells stained for NeuN do not show b-Gal signal for and EGFP fluorescence; only few cells show co-localisation of the tetO transgenes with neuronal marker. As it is, Fig.3 does not strongly support the conclusion of “widespread neuronal expression” and I have a doubt that quantitative analysis of correlation of b-Gal and EGFP signal with neuronal and glial markers will show predominance of their neuronal localisation. Authors themselves admit this implicitly, saying about strong EGFP expression in astrocytes and “low to moderate level” of EGFP expression in neurons (p.16597). Also, there is no reproducibility in the shape of cells stained for NeuN across the four columns in Fig.3C which might be an indication for the lack of specificity of antibodies.

2) Talking about reproducibility, the immunocytochemistry and EFGP data in Fig.3 are in striking controversy with previous data shown by the P. Haydon’s group (Pascual 2005; Halassa e2009, Lalo 2014). We observed more than hundred of brain slices from dn-SNARE mice and dnSNARE-negative littermates and did not see any neuronal EGFP fluorescence like that one shown in Fig3. Also, the pattern on astroglial EGFP-signals shown in Figure 3C (mainly diffuse staining with rare bright somata) is different from previous reports, showing numerous bright somata with bright bushy processes. Comparing EGFP-images shown by Fujita et al. with previous reports, one might wonder whether these images were obtained in the same strain of mice at all. The possible explanations of the discrepancy and reported “leakiness” of GFAP-tTA promoter might be a founder effect in the cohort of mice used by Fujita et al. or influence of epigenetic factors. It was reported that level of DNA methylation can strongly affect the GFAP expression in neurons and astrocytes (Barresi 1999; Hatada PLoS One 2008; Takizawa Dev Cell 2001). Again, one might only wonder why Authors imported the dn-SNARE mice not from P. Haydon’s lab, but from Ken McCarthy’s who did not publish any major paper on dn-SNAREs apart from his earlier collaboration with P. Haydon.

3) Analysis of basal leakiness of tet-Off system in Figure 3E may be misleading since qPCR data on transgenes expression are compared to the “clear” wild-type mice. No wonder that such comparison showed large, statistically significant difference. However, in physiological and behavioural experiments data recorded in dnSNARE Dox-Off mice are usually compared to the data from dnSNARE Dox-On mice or from the dnSNARE-negative littermates (Pascual et al. 2005; Halassa et al. 2009; Lalo et al. 2014). When compared in such way, the data in Fig.3E shows a different picture. Firstly, expression of dnSNARE transgene in Dox-On and GFAP-tTA/dnSNARE negative mice is much less than in Dox-Off. Secondly, expression of dnSNARE transgene in GFAP-tTA/dnSNARE negative mice is much less than Lac-Z and EGFP transgenes. Also, one can see from Fig.3C that characteristic astrocytic EGFP staining patterns disappear in the Dox-On and GFAP-tTA/dnSNARE negative mice. Combined, these data suggest that putative leakiness of tet-Off system under GFAP promoter affects mainly basal neuronal expression of transgenes whereas level of glial dnSNARE expression is much less in non-dnSNARE mice as compared to dnSNARE-Dox Off. So, even if GFAP promoter can be leaky, it can be mitigated by using the dnSNARE-negative littermates as a control.

4) Putting aside putative explanations of problems with cohort of dn-SNARE mice used by Authors, the data shown in Figures 2 and 3 are in some agreement that neuronal dnSNARE expression does not prevail even in that particular cohort. As shown in Fig.2E and stated in the Discussion, the cortical neuronal dn-SNARE expression reaches only 32% of glial level in 8-day old mice which is hardly big news since it is widely known that GFAP promoter can also be active in neurons as this stage of development. Furthermore, level of neuronal dn-SNARE expression was much less in the adult age. Although Authors do not provide a direct comparison of neuronal and glial fractions, estimation could be done using neuronal marker Rbfox3 as a reference. Comparing Figs.2 G-G, one could estimate the fraction of dn-SNARE expressing neurons as 1/16, i.e 8%. Even taking 20% as an “optimistic” estimate, one cannot draw a definitive conclusion about predominantly neuronal expression of dn-SNARE transgene in the adult mice without direct evidence of impairment of neurotransmitter release. Such evidence could be provided by demonstrating the significant decrease in the frequency of mIPSCs or mEPSCs or at least the decrease in the baseline fEPSPs in the neurons of off-DOX mice.

5) So, the paper lacks crucial evidence of “the profound suppression of synaptic transmission in Off-Dox dnSNARE mice”. In contrast, previous works on dnSNARE mice showed up-regulation of excitatory synaptic transmission in hippocampus (Pascual 2005) and up-regulation of inhibitory synaptic transmission in the neocortex (Lalo 2014) of dnSNARE mice. Thus, the dnSNARE mice (at least used in the previous work) do not have a major deficit in presynaptic release of neurotransmitters, strongly arguing against abundant neuronal dnSNARE expression. This inconvenient fact was ignored.

Authors show that removal of Dox induces the suppression of the EEG signal amplitude but attribute this effect to the neuronal expression of dnSNARE. Since Authors do no provide a convincing evidence of predominant dnSNARE expression in neurons as compared to glia (and show some evidence of the opposite) and do not provide any direct evidence of impaired synaptic signalling in neurons, this conclusion is unconvincing if not incorrect. One could argue that suppression of the EEG signal might be due to profound astrocytic dnSNARE expression, which was also observed in the present paper, causing an impairment of gliotransmitter release and changes in glial modulation of synaptic transmission. Actually, the up-regulation of GABAergic synaptic transmission, that we observed in dnSNARE mice (Lalo 2014) is an good agreement with this hypothesis.

6) Lastly, I would like to stress that in the former article we provide a several lines of evidence of exocytotic gliotransmission which cannot be affected by putative neuronal expression of dnSNARE transgene at all, in particular “sniffer-cell” detection of ATP release from isolated astrocytes and perfusion of individual astrocytes in situ

On 2015 Feb 11, Radboudumc Psycho-Oncology Journal Club commented:

A well-written, very informative and interesting paper with good clinical value. We believe it’s important to overcome barriers to accepting psychosocial services; not only for distress but other mental health problems as well. During the plenary discussion of this paper, our Journal Club generated the following comments:

1) 3070 patients completed the QUICATOUCH assessment of whom 10% scored above the threshold for distress (n = 310). We know from literature that psychological distress affects around 30% of all cancer survivors*. This percentage is substantially higher than the 10% found in current sample. We are wondering if there’s an explanation for this finding.

2) In the results section, table 1. Four of the big five cancer types (and Lymphoma are mentioned as seperate categories, complemented by a category named “other”. Because this “other” category is the largest one in actual number of patients, it would be interesting to know which cancer types it represents.

3) In the results section, we believe that the percentages for ‘Prefer to manage myself’, as portrayed in figure 1, do not match with the percentages mentioned in text. The figure displays a higher % of patients with a self-management preference for distress score 8-10 (53%) than distress score 4-5 (43%). In text, it’s the other way round.

*Mehnert A, 2014,Mitchell AJ, 2007,Zabora J, 2001

On 2015 Jun 02, Pierre Lindenbaum commented:

This article has been retracted (via Antoine Blanchard ‏@Enroweb ). An Expression of Concern has been published for this article "Expression of Concern on LaCour and Green, Science 346 (6215) 1366-1369." http://www.sciencemag.org/content/early/2015/05/20/science.aac6184

On 2015 Feb 15, William Grant commented:

The effect of diet on risk of Alzheimer's disease is very strong; vitamin D also reduces risk

The paper by Deckers and colleagues used the Delphi consensus approach to rank the target risk factors for dementia [1]. Diet was ranked 6th out of 11 factors in the Delphi ranking round 2, behind depression, diabetes, cognitive ability, physical activity, and midlife hypertension.

Ref. 1 relied primarily on prospective observational studies. While such studies are valuable, they are not the only way to assess risk. Another way is through ecological studies, either geographical or temporal. The first study to identify diet as an important risk factor for Alzheimer's disease was a multi-country ecological study [2]. It was discussed in two subsequent papers [3,4]. The most dramatic application of the ecological approach to studying the effect of diet on risk of Alzheimer's disease was in analyzing trends of Alzheimer's disease rates in Japan. The increase in Alzheimer's disease rates for elderly people in Japan from 1% in 1985 to 7% in 2008 was attributed to the nutrition transition from the traditional Japanese diet to the Western diet [5]. In addition, dietary advanced glycation end products from food cooked at high temperatures or aged as in hard cheese have been identified as an important risk factor for Alzheimer's disease [6, 7]. Diet is also a well known risk factor for diabetes.

A factor not mentioned but one that is gaining acceptance is low 25-hydroxyvitamin D [25(OH)D] concentrations [8]. Low 25(OH)D concentrations have been reported as a risk factor for Alzheimer's disease and dementia [9], diabetes mellitus [10], and cognitive dysfunction [11].

References 1. Deckers K, van Boxtel MP, Schiepers OJ, de Vugt M, Muñoz Sánchez JL, Anstey KJ, Brayne C, Dartigues JF, Engedal K, Kivipelto M, Ritchie K, Starr JM, Yaffe K, Irving K, Verhey FR, Köhler S. Target risk factors for dementia prevention: a systematic review and Delphi consensus study on the evidence from observational studies. Int J Geriatr Psychiatry. 2015;30(3):234-46. 2. Grant WB. Dietary links to Alzheimer's disease. Alz Dis Rev. 1997;2:42-55. http://www.sunarc.org/JAD97.pdf) 3. Grant WB. Dietary links to Alzheimer’s disease: 1999 update. J Alz Dis. 1999;1197-201. 4. Grant WB, Campbell A, Itzhaki RF, Savory J, The significance of environmental factors in the etiology of Alzheimer’s disease. J Alz Dis. 2002;4:179-89. 5. Grant WB. Trends in diet and Alzheimer’s disease during the nutrition transition in Japan and developing countries. J Alzheimers Dis. 2014;38(3):611-20. 6. Cai W, Uribarri J, Zhu L, Chen X, Swamy S, Zhao Z, Grosjean F, Simonaro C, Kuchel GA, Schnaider-Beeri M, Woodward M, Striker GE, Vlassara H. Oral glycotoxins are a modifiable cause of dementia and the metabolic syndrome in mice and humans. Proc Natl Acad Sci U S A. 2014;111:4940-5. 7. Perrone L, Grant WB. Observational and ecological studies of dietary advanced glycation end products in national diets and Alzheimer’s disease incidence and prevalence. J Alzheimers Dis. 2015 Jan 29. [Epub ahead of print] 8. Grant WB. Does vitamin D reduce the risk of dementia? J Alzheimers Dis. 2009;17(1):151-9. 9. Littlejohns TJ, Henley WE, Lang IA, Annweiler C, Beauchet O, Chaves PH, Fried L, Kestenbaum BR, Kuller LH, Langa KM, Lopez OL, Kos K, Soni M, Llewellyn DJ. Vitamin D and the risk of dementia and Alzheimer disease. Neurology. 2014;83(10):920-8. 10. Song Y, Wang L, Pittas AG, Del Gobbo LC, Zhang C, Manson JE, Hu FB. Blood 25-hydroxy vitamin D levels and incident type 2 diabetes: a meta-analysis of prospective studies. Diabetes Care. 2013;36(5):1422-8. 11. Annweiler C, Dursun E, Féron F, Gezen-Ak D, Kalueff AV, Littlejohns T, Llewellyn DJ, Millet P, Scott T, Tucker KL, Yilmazer S, Beauchet O. 'Vitamin D and cognition in older adults': updated international recommendations. J Intern Med. 2015;277(1):45-57.

On 2015 Jun 01, Attila Csordas commented:

associated dataset available via PRIDE/ProteomeXchange: http://www.ebi.ac.uk/pride/archive/projects/PXD001205

On 2015 Jan 12, Larry Parnell commented:

This paper was my selection for Paper of the Week for 8-12 Dec 2014.

On 2015 Jun 29, Bill Cayley commented:

A good example of when Less may be More: https://lessismoreebm.wordpress.com/2015/06/01/use-and-outcomes-associated-with-bridging-during-anticoagulation-interruptions-in-patients-with-atrial-fibrillation-findings-from-the-outcomes-registry-for-better-informed-treatment-of-atrial-fibrill/

On 2015 Jan 11, Donald Forsdyke commented:

CONTINUING SUPPORT FOR GRANTHAM'S GENOME HYPOTHESIS

Richard Grantham (1980), after examination of a mere 160 short sequences, proposed his ‘genome hypothesis.’ For nucleic acids he envisaged ‘manifold constraints and adaptations, of both structural and functional natures.’ These ‘could exist, independently of protein coding.’ Thus, there was ‘protein-independent molecular evolution of a non-neutral character.’ The distinctive ‘coding strategy of an organism’ was ‘at the heart of the problem of molecular evolution,’ and was likely to prove of fundamental importance for ‘speciation and systematics in general’ (Grantham et al. 1986).

These observations won extensive support when thousands of much longer sequences became available, even though many were incomplete (Forsdyke and Mortimer 2000; Mortimer and Forsdyke 2003; Lee et al. 2004). The data, consistent with the early base compositional studies of Chargaff, Sueoka and Szybalski, are now further affirmed by this comprehensive new work involving thousands of entirely complete sequences (Goncearenco and Berezovsky 2014).

However, while there is little disagreement on data and the need for evolutionary trade-offs (‘mutual adjustment of the nucleotide and amino acid compositions’), readers should note that there remains disagreement over interpretations (see my comments on a previous paper and my textbook; Zeldovich et al. 2007, Forsdyke 2011). Readers should also note that the present text (p. 3) has proline (P) listed in both the high GC% saturation group and the low GC% saturation group. In error, phenylalanine (F) was replaced by P in the latter. The GC% low and medium groups have amino acids listed in order of increasing codon GC% saturation, but the GC% high group has amino acids listed (p. 3) in decreasing order of codon GC% saturation (see Fig. S3). And it is puzzling that Figures 1a and 1b appear the same, with just axis labels interchanged, yet some points seem incorrectly interchanged. The rectilinear interpretations of obvious curvilinear relationships (Figs. 8, S9) are also problematic, as noted by Reviewer 2.

The notion that aspartate and glutamate (with purine-rich codons GAY and GAR) ‘cannot be used for the efficient tuning of the nucleotide composition,’ does not hold for the tuning of AG%. Furthermore, it should be noted that as GC% increases, the decline of A and T does not affect both bases equally. While G% and T% tend to remain constant, C increases at the expense of A. Likewise, when GC% decreases, A increases at the expense of C (Mortimer and Forsdyke 2003). This A-for-C transversional trading, most evident at extreme GC% values (Fig. S10), should decrease the probabilities of G-quadruplexes and thymine dimers. Finally, noting for example the high AG% in thermophiles, it may be premature to conclude that tradeoffs are a ‘purely compositional phenomenon, linking the realms of nucleic and amino acids in prokaryotes regardless of their life styles, environments, and phylogeny.’ Grantham’s admonition regarding speciation and systematics should not go unheeded.

Forsdyke DR: Evolutionary Bioinformatics. 2nd edition. New York: Springer, 2011.

Forsdyke DR, Mortimer JR: Chargaff’s legacy. Gene 2000, 261:127-137.Forsdyke DR, 2000

Goncearenco A, Berezovsky IN: The fundamental trade-off in genomes and proteomes of prokaryotes established by the genetic code, codon entropy, and physics of nucleic acids and proteins. Biology Direct 2014, 9:29.Goncearenco A, 2014

Grantham R: Workings of the genetic code. Trends Biochem Sci 1980, 5:327-331.

Grantham R, Perrin P, Mouchiroud D: Patterns in codon usage of different kinds of species. Oxford Surv Evol Biol 1986, 3:48-81.

Lee S-J, Mortimer JR, Forsdyke DR: Genomic conflict settled in favour of the species rather than of the gene at extreme GC% values. Applied Bioinformatics 2004, 3:219-228.Lee SJ, 2004

Mortimer JR, Forsdyke DR: Comparison of responses by bacteriophage and bacteria to pressures on the base composition of open reading frames. Applied Bioinformatics 2003, 2:47-62.Mortimer JR, 2003

Zeldovich KB, Berezovsky IN, Shakhnovich EI: Protein and DNA sequence determinants of thermophilic adaptation. PLoS Comput Biol 2007, 3(1):e5.Zeldovich KB, 2007

On 2015 Mar 29, Geriatric Medicine Journal Club commented:

The full version of study of the American Geriatrics Society Clinical Practice Guideline for Postoperative Delirium in Older Adults was reviewed at the March 2015 Geriatric Medicine Journal Club (follow #GeriMedJC on Twitter). The full discussion can be found at: http://gerimedjc.blogspot.com/2015/03/gerimedjc-march-27-2015.html?spref=tw This is an important document as surgical specialists identified delirium as “essential" more than any other topic in the care of older adults. While the #GeriMedJC discussion noted there were no real surprises in this data synthesis, hopefully this guideline will serve to fill a knowledge gap.

On 2014 Dec 13, David Keller commented:

The grandfathered internists had 20% more experience in practice than the MOC'ed internists

The "grandfathered" internists in this study were certified in 1989, while the "MOC'ed" internists (those forced to undergo Maintenance of Certification) were certified in 1991. The study was conducted in 2001, so the grandfathered internists had 12 years in practice, while the MOC'ed internists had only 10 years in practice. The difference of 20 percent more practice experience could account for much of the difference between the two groups; for example, the fact that the more experienced internists spent more money on patient care might reflect their increased experience in billing Medicare and insurance companies, which is knowledge not taught during the MOC process; rather, skill at billing is acquired through years in practice.

On 2016 May 23, Heidi Schulz commented:

The p.T4I mutation had been previously published by Kinnick et al. 2011 (PMID21273940).

On 2015 Jan 20, Donald Forsdyke commented:

HYPOTHESES OF SPECIATION As hinted at by one of the reviewers, hypotheses on the initiation of speciation are broadly categorized as genic and chromosomal (Nevo, 2012). The possibility that their intriguing observations on putatively sterile hybrids between certain mouse subspecies might be explained in chromosomal (non-genic) terms is not considered by Turner and Harr. Yet, from studies of hybrids of the same subspecies, Bhattacharyya et al. (2013; 2014) infer that “meiotic asynapsis of heterospecific homologous chromosomes is the primary mechanistic basis of hybrid sterility.” This indicates a role for “a fast-evolving subset of the noncoding genomic sequence important for chromosome pairing and synapsis.” Thus, any observed genic differences would be secondary to this (Page and Orr-Weaver, 1997).

Furthermore, it is incorrectly implied by Turner and Harr that the work of Bateson (1909) supports the genic viewpoint to which the names of Dobzhansky and Muller are attached (“DM incompatibilities”). This is not a minor point, since Bateson consistently favored a non-genic viewpoint that is today best equated with the chromosomal hypothesis (Nei and Nozawa, 2011; Forsdyke, 2011).

Bhattacharyya T, Gregorova S, Mihola O, Anger M, Sebestova J, Denny P, Simecek P, Forejt J. 2013. Mechanistic basis of infertility of mouse intersubspecific hybrids. Proceedings of the National Academy of Sciences, U S A 110: E468–477. doi:10.1073/pnas.1219126110.

Bhattacharyya T, Reifova R, Gregorova S, Simecek P, Gergelits V, Mistrik M, Martincova I, Pialek J, Forejt J. 2014. X chromosome control of meiotic chromosome synapsis in mouse inter-subspecific hybrids. PLoS Genetics 10: e1004088. doi:10.1371/journal.pgen.1004088

Forsdyke DR. 2011. The ‘B’ in BDM. William Bateson did not advocate a genic speciation theory. Heredity 106:202. doi:10.1038/hdy.2010.15.

Nei M, Nozawa M. 2011. Roles of mutation and selection in speciation: from Hugo de Vries to the modern genomic era. Genome Biology and Evolution 3,812–829. doi:10.1093/gbe/evr028.

Nevo E. 2012. Speciation: chromosomal mechanisms. In: eLS. Chichester: John Wiley & Sons. doi: 10.1002/9780470015902.a0001757.pub3.

Page AW, Orr-Weaver TL. 1997. Stopping and starting the meiotic cycle. Current Opinion in Genetics and Development 7:23–31. doi: 10.1016/S0959-437X(97)80105-0.

On 2014 Dec 20, Rafael Najmanovich commented:

Readers of this article may find interesting our large-scale rare codon cluster analysis across Pfam protein families (Chartier M, 2012).

On 2016 Jun 29, Bill Noble commented:

Hi Zhibin, thank you for your comment. Yes, as we describe on p. 1148 in the paragraph beginning "Our primary goal in this article ...", there are many methods described in the literature for obtaining approximately calibrated scores. These are variously referred to as E-values or p-values, though the precise semantics of any such score depends on the method used to generate it. One of the main points of our article is to introduce an independent method to help evaluate the calibration properties of these types of scores. For example, Figure 3 shows the calibration properties of the MS-GF+ E-value. --Bill Noble & Uri Keich

On 2016 Jun 24, Zhibin Ning commented:

Actually, quite a few search engine and downstream workflows do not use "score" to rank the peptide, instead, p-value is more appropriate to do the "second job" claimed in this manuscript. P value (or E value), in some sense, is the calibrated "score".

On 2014 Dec 11, David Keller commented:

Congress gave health insurers the right to violate patient privacy - this must end

Physicians insert “boilerplate text", defined as words which convey no meaningful medical information or insights, into their clinical notes mainly to satisfy documentation requirements imposed by the Medicare Evaluation and Management (E&M) Services Guide. Medicare and nearly all private health insurers adhere to the E&M Guide to determine their payments to physicians. These E&M service definitions largely dictate the format and content of physicians’ progress notes.

There is a natural tension between physicians, who want to be paid in full for their efforts, and payers, who want to minimize payments to physicians. Payers enforce physician adherence to E&M documentation and billing regulations by hiring clerical employees called “coders” to audit patients’ clinical charts for documentation deficiencies. To avoid fines and prosecution, physician notes are usually generated first and foremost to fulfill the requirements of these coder audits, rather than the needs of their fellow clinicians.

Coders determine whether a progress note meets the billing requirements mainly by checking how many elements of defined data are present. Most physicians are not completely familiar with the complex, confusing, and arcane E&M requirements, so they load up their notes with as much computer-generated boilerplate as they can, hoping it will include an overlooked element of data needed to satisfy a coder’s audit.

This whole system must be discarded, or there can never be any hope of having progress notes that communicate clearly and concisely between clinicians. A quick and sure way to end this “tyranny of the coders” is for Congress to terminate the exemption granted to payers by the patient privacy laws, including HIPAA. This would eliminate inspection of the clinical chart by payers and their coders. Payers would no longer need battalions of coders, who could be relieved of their duties and retrained to perform more productive tasks.

Payers should not be allowed to read, inspect, or audit patients’ clinical charts; this is a violation of patients’ medical privacy. It harms patients by causing physicians to spend more of their limited time generating progress notes, leaving less time for interacting with patients. Payers must verify physician billings in some way which does not violate the privacy of patient charts. Payments could be determined by the average amount of physician time required to treat the patient’s illnesses and manage their chronic conditions, as determined from the submitted diagnosis codes. Physicians could submit attestations for any required variations. In a world without E&M coding requirements, the clarity and efficiency of clinical documentation could be improved enormously, with more time for physicians to actually examine and treat patients.

On 2017 Jul 13, Randi Pechacek commented:

Holly Ganz mentioned this paper on microBEnet as part of a larger discussion on antibiotic use in raising livestock.

On 2014 Dec 12, David Keller commented:

Parkinsonian constipation & altered gut microbiota: suggestions for further study

It has long been known that constipation due to autonomic dysfunction is an early pre-motor finding in Parkinson disease (PD). Other bowel motility disorders have been causally linked with alterations in gut flora; for example, opioid-induced constipation is a known risk factor for C. Difficile overgrowth (1,2), which in turn can cause diarrhea. If Parkinsonian constipation promotes changes in gut flora, it is reasonable to ask whether these changes in gut flora might alter the course of PD itself.

Do toxins produced by gut microbes contribute to CNS neurodegeneration in susceptible individuals? The serum levels of these toxins should be tracked in a study designed to measure their effects in PD. If overgrowth of certain toxigenic bacterial species appears to cause accelerated neurodegeneration, then treatment with antibiotics or probiotics (or both) should be tested in selected individuals, in an attempt to slow the progression of their PD.

References

1: Keller DL. Opioid use and clostridium difficile infection. Am J Med. 2013 Apr;126(4):e13. doi: 10.1016/j.amjmed.2012.08.024. PubMed PMID: 23507210.

2: Mora AL, Salazar M, Pablo-Caeiro J, Frost CP, Yadav Y, DuPont HL, Garey KW. Moderate to high use of opioid analgesics are associated with an increased risk of Clostridium difficile infection. Am J Med Sci. 2012 Apr;343(4):277-80. doi: 10.1097/MAJ.0b013e31822f42eb. PubMed PMID: 21934595.

On 2017 Mar 23, University of Kansas School of Nursing Journal Club commented:

Team Members: Jason Tracy, Abigail Spare, Julia Damman, Laura Thompson, Sophie Bono, Maeve Markey. [Class of 2017]

Background

In our class, we took time to study shared governance in acute care organizations and looked at the importance of shared governance in regards to all nursing positions. This article not only touches on the importance of shared governance, but also explores the processes needed to help direct patient care nurses transition their knowledge, skills, and attitudes into the leadership role that would facilitate shared governance. Our team was especially interested in the idea of shared governance, and liked how this particular article discussed an intervention to increase shared governance by empowering nurses with leadership. This is an interesting, practical intervention that could be implemented with nurses.

Methods

This was a descriptive qualitative study that took place at a university medical center hospital with 300 adult beds in southeastern U.S. Through partnership, the CNO and CON faculty developed a Frontline Innovations (FI) group that help mentor frontline nurses with leadership skills, problem solving strategies, and evidence based to fix identified operational failures. The group included 12 frontline nurses from 4 different medical surgical units chosen by the CNO based on their leadership potential. Frontline nurses are extremely important to include in change process because they have direct impact to patients outcomes.

Interviews were conducted to reveal the steps of process improvement, and analysis of interviews determine recurring themes which were divided into process and outcome measures describing staff, administration, and faculty interaction. Engagement was identified as a component of the process theme and was observed through both verbal and nonverbal communication with the CNO, as well as focused role playing activities with interprofessional groups. The nurses that engaged in these group activities were then responsible for sharing what they learned with their peers on the unit (Dearmon et al. 2015).

The study ultimately affected nursing administration and executives as well because of their influence in the structures and processes of various units. This influence carried over to the care delivery of frontline nurses give. Shared governance distributes power, resources, and information as equally as possible, so it is important that both administration and frontline nurses work together (Dearmon et al. 2015).

Findings

Outcome measures included collaboration, empowerment, confidence, and lifelong learning. Collaboration was found to be important for the creation of the organization’s new shared governance model. It allowed members to work together to achieve a common goal, strengthen group morale, and improve the credibility of decisions making process. Nurses began to feel empowered when they felt their voices were heard by the CNO and changes were made. Nurses who were given the opportunity to visit other facilities that has shared governance model returned to their own organization feeling empowered to brought back shared governance model in their unit. In order to instill confidence that the new shared governance model would work for the organization, Frontline Innovations group members coached nurses through the new project and allowed them to use the process to see success for themselves (Dearmon et al. 2015).

Creating the Frontline Innovations group proved to be an effective way to influence change at the bedside through the implementation of a shared governance model. All disciplines involved learned to trust each other and this transferred to practice, improving collaboration in patient care. Nurses became more enthusiastic about the care they provided because they felt their voices were being heard and thus began to realize they had the ability to change practice when gaps were seen. The transition from a culture of administrative decision-making to one of staff engagement and shared decision-making forged an organization with all members willing to take responsibility for workplace obstacles (Dearmon et al. 2015).

Nursing Implications

The healthcare industry continues to become more complex as new discoveries are made. As these complexities expand the nursing practice becomes more specialized, making it more difficult for the executives in charge to understand the needs of the individual environments. Only by including the nurses that work in direct patient care areas can executives get a full picture of the operations and needs of the units. The literature shows that although both executive and front line nurses desire to collaborate, there is a gap to be filled before this can occur. The literature shows that there are means by which this gap can be bridged by developing mentorship programs, from which front line nurses can gain the knowledge needed to participate in counsels that effect change within the organization. This also helps nurses find their voice and the desire to work independently with their executive mentors to make quality improvements within their units. The end result was better working relationships between the two groups and more productive initiatives in providing better patient care (Dearmon et al. 2015). This means the healthcare industry is able to provide more efficient services and better patient outcomes due to changes made by the people involved in doing the work. As new nurses getting ready to enter the healthcare system, it is reassuring to know that our point of view is just as important as that of a nurse executive. Stepping into a new organization, we bring a fresh set of eyes that can see things to which others have become accustomed to. Being able to affect change within an organization is an important factor when choosing a career, and this article helps support this initiative. The section that discussed nurses being able to work independently gave rise to the idea of working and improving within specific system. Individually both systems can do a certain amount of work, but without collaboration they can ultimately fail. However, this also requires leadership to be in place in both systems. This means that although the executive officers may be the faces of the organization, nurse leaders in each system are called to step up to the task of effecting change.

If the literature is found to be valid and reliable, then many changes are sure to come. There is already a push for hospitals to obtain Magnet status. The evidence found in this article supports the Magnet program and the need for more nurses stepping into leadership roles. This may also change how nurses are integrated into an organization by mandating new employees to be trained in the organization's leadership development program. The end product would include more staff educated in how to effectively work within their unit’s structure and how their unit collaborates best with other units. This alone can provide a multitude of solutions to decrease expenses and the amount of waste, provide more efficient care, and increase employee morale.

Reference

Dearmon, V., Riley, H., Mestas, G., & Buckner, B. (2015). Bridge to shared governance: Developing leadership of frontline nurses. Nursing Administration Quarterly, 39(1), 69-77. doi: 10.1097/NAQ.0000000000000082

On 2014 Dec 05, David Mage commented:

This has been known for a long time as described by Sudden unexpected infant deaths in Dundee, 1882-1891: overlying or SIDS?

Williams FL, Lang GA, Mage DT.

Scott Med J. 2001 Apr;46(2):43-7.

PMID: 11394337

On 2015 Apr 02, University of Lausanne Genomics, Ecology and Evolution Journal Club commented:

The article nicely highlights the need for a reference genome panel consisting of African variants and understanding African population history. This is a summary of the discussion at our journal club. Gurdasani et al. offer evidence for B the antu expansion, although the sample distribution came mostly from the coast rather then inside of the African continent where the Bantu expansion occurred. Moreover, the article lacks discussion of the gene flow effects other then the impact of Euroasian admixture on the differentiation among the African populations. It would be interesting to see are some effects due to allele surfing or allele fixation. In the article, the highest proportion of Euroasian ancestry is observed in the Ethiopian population. The authors hypothesize that Euroasian ancestry is responsible for the differentiation in African populations, rather than adaptation to selective forces. But, the highly differentiated SNPs between Euroasian and African populations showed evidence of differentiation within genes of malaria susceptibility and osmoregulation, which makes the previous statement imprecise. Also, the results of population admixture shows an overlap both in Euroasian and HG ancestry in East Africa without any further explanation of how it happened. The last part of the article is probably the most valuable one, since it shows the improvement of imputation accuracy of African populations variants by using different reference genome panel. Although finding the optimal array design for capturing most of African variation is admirable, in the current state of sequencing technology it might not be necessary anymore. A final remark is that the visualization of the results in extended and supplementary data was sometimes hard to decipher.

On 2014 Dec 04, Rafael Najmanovich commented:

We predicted back in 2012 (Chartier M, 2012) as part of a large scale analysis of rare codon cluster that such clusters may play a role in molecular recognition whereby translational pauses in the nascent protein would permit its interaction with other proteins to permit intracellular targeting and membrane insertion. It is unfortunate that the authors were not aware of our publication.

On 2015 Jun 20, Arnaud Chiolero MD PhD commented:

Beliefs are pervasive in the epidemiology of nutrition, whose scientific credibility is seriously questioned. This paper offers some solutions to this major issue.

On 2014 Dec 10, Jim Woodgett commented:

There is no direct evidence in this paper that the effects of lithium on kidney function are specifically via GSK-3beta. The GSK-3alpha isoform, which was not investigated, is equally inhibited by this ion and the following statement is erroneous: "lithium, a selective inhibitor of GSK3β". As an aside, dominant negative GSK-3beta (kinase-dead), interferes with both GSK-3alpha and beta.

On 2015 Feb 06, Ryan Radecki commented:

Post-publication commentary: "More Futility for Mechanical CPR"

A few weeks ago we reported on the use of the LUCAS-2 for crushing the thorax, resulting in significant internal injuries. None of these injuries were judged to have contributed to any patient’s demise – but, still, concerning. This could be more forgivable if there were other advantages, say, survival.

Unfortunately, yet again, mechanical CPR fails to demonstrate superiority....

http://www.emlitofnote.com/2014/12/more-futility-for-mechanical-cpr.html

On 2016 May 25, Jonathan Rakar commented:

This paper is part of a dissertation (2016), available here: http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-125925

On 2015 Jan 22, Larry Parnell commented:

This paper was my selection for Paper of the Week for 12-16 Jan 2015.

We have it rather interesting to map onto the network described in this article (Fig 2) those genes carrying variants that associate with obesity anthropometrics. Some regions of the network are sparsely populated or entirely devoid of such obesity GWAS genes, while others have a great concentration of such genes.

We were surprised not to see FTO or extracellular remodeling genes/proteins displayed in this network. The FTO/IRX3 locus has much evidence linking it to obesity. As adipose expands, much remodeling of the extracellular matrix takes place.

On 2015 Jun 22, Ivan Oransky commented:

This letter has been retracted, for an unusual reason: http://retractionwatch.com/2015/06/22/mystery-writer-impersonates-cardiovascular-pathologist-and-pens-retracted-paper/

On 2015 Aug 10, NephJC - Nephrology Journal Club commented:

This study was discussed on July 28 and 29th 2015 in the open online nephrology journal club, #NephJC, on twitter .

Introductory explanatory comments, written by Matt Sparks, are available at the NephJC.

There was a very lively discussion, which included about 50 participants, enriched by the active engagement by two of the authors, Ben Humphreys and Rafael Kramann, and was complete with many citations to related research articles.

A transcript and a curated (i.e. Storified) version of the tweetchat are available from the NephJC website.

The highlights of the tweetchat discussion were:

• The study does help establish that Gli1+ perivascular ('pericytes') MSC-like cells are the source of ~70% of myofibroblasts after organ injury, and are key players in the subsequent fibrosis.

• Successful/effective targeting of this pathway leads to less fibrosis in preservation of heart function after transverse aortic banding (a heart failure model), however substantial hurdles still exist in terms of delivering a potential therapeutic intervention to lessen fibrosis but ensuring the protective effects of fibrosis are not inhibited.

• The authors' did remarkable work in leveraging elegant techniques such as lineage tracing, multiple organ injury models, parabiosis and diphtheria toxin/Gli1 knockout.

Interested individuals can track and join in the conversation by following @NephJC or #NephJC, or visit the webpage at NephJC.com.

On 2017 Jul 14, Randi Pechacek commented:

Amy Pruden, co-author of this paper, wrote an entertaining blog post on microBEnet providing background.

On 2016 Jul 01, Nicholas J L Brown commented:

Our reanalysis of the data from the article by Vedhara et al. (2015) showed that their results are fragile and ambiguous and depend on (a) the inclusion or exclusion of two study participants missing values for one of the control variables and (b) the accumulation of statistical suppression effects in the regression analyses. See 10.1016/j.paid.2015.11.055.

On 2014 Dec 23, GEANNCARLO LUGO commented:

This is a timely review article addressing the role of the STAT3-dependen signaling pathway in Tuberculosis. I find it easy to read, well structured, and quite helpful in summarizing everything that is known about these factor in the context of TB.

On 2016 Aug 05, Zvi Herzig commented:

Adding to Bates' comments below, here are some specific misrepresentations of the primary evidence by the Pisinger and Døssing review, relating to almost every major issue in e-cigarette toxicology (not an exhaustive list):

• “Some studies found high maximum concentrations of total TSNA”, citing studies showing TSNAs 100-200 times below cigarette smoke levels [1-3].
• “Exposure to formaldehyde was comparable with smoking", referring to a study calculating formaldehyde levels nine times below that from tobacco smoke [1].
• “Propylene glycol has been found to exacerbate and/or induce multiple allergic symptoms in children”, citing a study stating that "apparently… outcomes were not driven by propylene glycol” [4].
• “Values below the threshold limit don't necessarily protect against the health effect of 200–300 daily inhalations over decades”, referring to safety limits calculated for 8 hours exposures “day after day, over a working lifetime” [5].
• “These metals appear on the U.S. Food and Drug Administration's 'Harmful and Potentially Harmful Chemicals' list”, referring to metals detected below levels acceptable to the FDA for chronic inhalation [6-8].

[1] Goniewicz ML, Knysak J, Gawron M et al. Levels of selected carcinogens and toxicants in vapour from electronic cigarettes. Tob Control 2014

[2] Kim HJ, Shin HS. Determination of tobacco-specific nitrosamines in replacement liquids of electronic cigarettes by liquid chromatography-tandem mass spectrometry. J Chromatogr A 2013

[3] Farsalinos KE, Romagna G, Voudris V. Authors miss the opportunity to discuss important public health implications. J Chromatogr A 2013

[4] Choi H, Schmidbauer N, Spengler J et al. Sources of propylene glycol and glycol ethers in air at home. Int J Environ Res Public Health 2010

[5] ACGIH. Chemical Substances Introduction. acgih.org 2016. http://www.acgih.org/tlv-bei-guidelines/tlv-chemical-substances-introduction

[6] Siegel M. Metals in Electronic Cigarette Vapor are Below USP Standards for Metals in Inhalation Medications. The Rest of the Story: Tobacco News Analysis and Commentary. 2013.http://tobaccoanalysis.blogspot.com/2013/04/metals-in-electronic-cigarette-vapor.html .

[7] Farsalinos K. Metals and nanoparticles in e-cigarettes. 2013.http://www.ecigarette-research.com/web/index.php/2013-04-07-09-50-07/2013/89-metals-and-nanoparticles-i n-e-cigarettes-commentary.

[8] Farsalinos K, Voudris V, Poulas K. Are Metals Emitted from Electronic Cigarettes a Reason for Health Concern? A Risk-Assessment Analysis of Currently Available Literature. International Journal of Environmental Research and Public Health 2015

On 2016 Aug 05, Zvi Herzig commented:

None

On 2016 Aug 16, Clive Bates commented:

Note: The focus on conflicts of interests (COI) received criticism both for exaggeration and for the tacit assumption that a COI declaration somehow invalidates the scientific merit of the work. See Kosmider L, 2016 Ideology versus evidence: Investigating the claim that the literature on e-cigarettes is undermined by material conflict of interest and the reply from the authors Pisinger C, 2016 Reading the conflict of interest statement is as important as reading the result section: Response to the letter by Dr. Kosmider.

COI declaration is not intended as a form of self-invalidation, but for transparency. It is absurd to assert that it is as important as the substance of the work itself. COI tends to focus on conflicts arising from commercial interests, but many researchers in this field have undisclosed conflicts relating to funders, regulators, employers' prior policy positions, and their long-held beliefs. If these authors want to investigate COI (not the ostensible purpose of this work), then they should do it rigorously, and with a broader view of COI.

On 2016 Aug 05, Clive Bates commented:

Despite the title, the review does not address the "health effects of electronic cigarettes". Or perhaps to the extent that it does, it does not find anything negative to report. This may explain the diversionary conclusion, which resorts to an argument about competing interests rather than a critique of actual scientific findings [see note appended below].

It does, however, attempt to summarise various proxies for health effects but does this very poorly. Throughout the substance of the article, it fails to internalise the most basic concept in toxicology:

"the dose makes the poison" (attributed to Paracelsus, 1538).

The assessment of e-cigarette science by the Royal College of Physicians (London), Nicotine without smoke: tobacco harm reduction, 28 April 2016 included this review but passed over it in the following terms:

Recent reviews of the health effects of toxins inhaled during normal use of e-cigarettes have expressed concerns over potential adverse effects based on the presence of these contaminants, but not their levels, which are generally the more important determinant of toxicity.

The danger of a review that approaches toxicology in this way is that plays into public fears inspired by chemical names and may deter smokers, who are at well-established serious risk from high exposures to harmful agents, from switching to the much safer vapour products. In doing so, the careless framing of scientific findings can cause actual harm.

In addition, such work provides a citable source for irresponsible activism, and so can entrench misunderstanding in regulatory bureaucracies like the WHO, FDA and European Commission - further amplifying the harm through ill-conceived regulation.

For any future work, these authors or others should carefully distinguish between the following:

• The detection of the presence of a hazardous agent in vapour, recognising a de minimis level for which there is no material concern - see Burstyn I, 2014 for an approach that adopts occupational exposures as a frame of reference.

• Human exposure to a hazardous agent that may or may not present a risk to health recognising that magnitude of exposure and materiality matter. For an insight into how to approach this, see Farsalinos KE, 2014. These authors provide a good example in Table 3, showing e-cigarette nitrosamine exposure to be about 1,000 times lower than in cigarette smoke and comparable to the residual levels found in regulated medicinal NRT.

• Observed damage to cells found in vitro following exposure to vapour that may have no bearing on human exposure or risk. See criticism of the extreme misinterpretation of one cell study finding in PubMed Commons on Yu V, 2016. This study provides a good example of what not to do.

• Observed impacts in animal studies, which are rarely a reliable proxy for human experience - see discussion by the Laura and John Arnold Foundation of the weakness of animal studies: Why journalists should stop publishing mouse studies

• The illusion of risks induced by operating vaping equipment in a way that no human user will ever experience other than momentarily. The Jensen RP, 2015 study on 'hidden formaldehyde' is a case in point and has been severely and appropriately criticised (Bates CD, 2015) for running equipment too hot and in 'dry puff' conditions, and then sensationalising the findings as if they showed a cancer risk, and suggesting these may be greater than for smoking.

• Observed changes to the human body that are not necessarily an indication of harm to health. Nicotine and coffee both create changes to the cardiovascular system, but neither is a direct cause of material harm - see PubMed Commons review of Carnevale R, 2016.

• Signs of immediate health gains when e-cigarettes are used as an alternative to smoking - for examples see Polosa R, 2015 and Polosa R, 2014 for examples of 'harm reversal' in chronic respiratory patients.

Finally, given that almost all e-cigarette users are former or current smokers, then any attempt to provide useful information on e-cigarette risk should always use cigarette smoking and related risk as a point of reference, not only the risks compared to complete abstinence. To avoid this comparison is to miss the value of harm reduction as a strategy in tobacco control. There are many examples of this failure - to take just four:

• Allen JG, 2016 raised the alarm about an additive, diacetyl, but failed to point out that, as one critic put it, "average diacetyl exposure from vaping is 750 times lower than from smoking".

• Yu V, 2016 despite concluding that "e-cigarette vapor, both with and without nicotine, is cytotoxic to epithelial cell lines and is a DNA strand break-inducing agent", the study showed that cells exposed to the e-cigarette vapour medium were still alive after 8 weeks but all were dead in the tobacco smoke extract within 24 hours. This important comparison was buried in the paper, not mentioned in the abstract, and not included in the authors' publicity statements about the study.

• Schweitzer KS, 2015 attempted to show that nicotine causes "dose-dependent loss of lung endothelial barrier function, which is associated with oxidative stress and brisk inflammation". But this study has been criticised for using nicotine exposures 500-50,000 times higher than nicotine concentrations in the blood of typical smokers or vapers: Cell studies on e-cigarettes: don’t waste your time reading (at least most of) them

• Paley GL, 2016 concluded that e-cigarette battery fires and explosions were a "significant public health risk", based on six reported incidents. The authors asserted this without noting that in 2011 there were 90,000 smoking-material fires in the U.S. resulting in an estimated 540 civilian deaths, 1,640 civilian injuries and $621 million in direct property damage. To the extent vaping replaces smoking it will reduce this toll and result in a significant public health win.

On 2015 Jan 25, Reuben Strayer commented:

If anyone wants the full (not truncated) version of the manuscript, contact me at the email address above.

On 2017 Jul 09, Jeffrey Ross-Ibarra commented:

In our manuscript exploring the population genetics of local adaptation (Tiffin and Ross-Ibarra 2014) we included a discussion about the potential uses of reduced representation data (e.g. RAD-seq, GBS). To provide a sense of the probability of using reduced representation data to identify targets of selection, we included a figure showing the probability of having a SNP included in a region of the genome in which diversity had been severely reduced due to a recent selective sweep. Unfortunately this figure is not correct; an error in the code inadvertently used centimorgans as morgans, causing the recombination rate to be off by a factor of 100.

To correct this we have generated a new figure (see http://rpubs.com/rossibarra/257207; raw code is available at https://gist.github.com/rossibarra/be44cc3b3796f45840d942ad11c01ba1) that corrects this error and presents a more realistic model. Our previous model assumed SNPs were distributed evenly across the genome and the presence of a single SNP near a sweep was sufficient for detection. Instead, here we explicitly model sequence “tags” coming from RAD-seq or GBS, and incorporate information about the variation in diversity expected among tags in neutral regions of the genome. The figure clearly shows that with dense marker coverage and strong selection, the probability of detecting reductions in diversity due to recent selective sweeps from new beneficial mutations can be relatively high. We emphasize, however, that the purpose of the figure is solely to develop an intuition of the likelihood of detecting a recent selective sweep. The many simplifying assumptions made in generating the figure (no recent demographic change, both sequence tags and recombination occur uniformly along the genome, selection is on a novel beneficial mutation with additive effect that has recently swept to fixation), as well as the specific mutation rates, sample size, sequence length, and recombination rates assumed will all affect the actual probability of a tag being included in a selective sweep. Moreover, this figure does not touch on many other relevant issue such as multiple testing, complex demography, background selection, or other modes of positive selection (e.g. from standing variation, balancing selection, or selection on polygenic traits).

We have submitted a correction to the journal.

We thank Eric Johnson for drawing our attention to the error, and Eric Johnson, Kathleen Lotterhos, and Graham Coop for kindly reviewing previous versions of the code and assumptions we have used in generating this new figure.

On 2014 Dec 03, Paolo Tieri commented:

This article is one of the "Multi-omic Data Integration" Research Topic in Frontiers, which focuses on data integration approaches and methods of any type and extent, their application in understanding the pathogenesis of specific diseases or in identifying candidate biomarkers, in order to exploit the full benefit of multi-omic datasets and their intrinsic information content. For more information about multi-omic data integration check http://journal.frontiersin.org/ResearchTopic/2280

On 2015 Mar 17, Alan Taylor commented:

This confirms work done to determine causes of OA in elderly patients. Aromatase deficiency is a key contributor to OA. Not only is it responsible for the final step in the synthesis of estrogen from androstendione but also converts testosterone to estrogen in males. The estrogen acts on two E receptors in female bone cells/ in males there is one androgen receptor and one estrogen receptor. These are a essential for bone health and maintenance. In the genome wide study under expression of CYP19 gene that codes for aromatase was evident in most cases. Severe OA in a male has been successfully treated for 15 years using a low dose of transdermal estrogen. This effect has been shown to be reversible It takes 6 to 8 weeks for the treatment to reverse the OA. Discontinuation causes OA to redevelop after 6-8 weeks. Recommencement again reverses the OA to give no symptoms whatsoever. Aromatase inhibitors will obviously prevent the two E receptors in the osteocytes from working. In males testosterone deficiency also will prevent the aromatase from catalysing the formation of estrogen.

On 2015 Apr 12, James Tsung commented:

Video of Right Ventricular Puncture During Pericardiocentesis for Cardiac Tamponade: https://youtu.be/WflPJCDl34I

On 2014 Dec 15, Sean Ekins commented:

Also added in my roundup of 2014 papers http://www.collabchem.com/2014/12/15/a-year-in-publications-2014/

On 2014 Dec 04, Sean Ekins commented:

Here is a link to a discussion of how this came to be published in DDT and not a stem cell journal http://www.collabchem.com/2014/12/04/chemical-probes-and-parallel-database-worlds-who-wants-to-know-more-publishing-fun/

On 2015 Feb 05, G Gasim commented:

Dear Dr.Yousif congratulations for the great efforts

do you have an explanation for the variation in genotypes between monoinfected Sudanese patients and those coinfected with HIV?

On 2015 Apr 26, Geriatric Medicine Journal Club commented:

This is a systematic review of non-pharmacologic interventions for orthostatic hypotension including exercise, FES, compression, physical countermaneuvers, head of bed up, water intake, and meal strategies. This article was critically appraised at the April 2015 Geriatric Medicine Journal Club (follow #GeriMedJC on Twitter). The full discussion can be found at: http://gerimedjc.blogspot.com/2015/04/april-2015-gerimedjc.html?spref=tw An interesting finding was that an acute bout of exercise may exacerbate orthostatic hypotension in short term. This review did not cover interventions like salt intake.

On 2014 Dec 17, Sean Ekins commented:

Some slides on earlier work are here http://www.slideshare.net/ekinssean/applying-computational-models-for-transporters-to-predict-toxicity

On 2015 Feb 09, David Keller commented:

Dr. Lee,

Your arguments against prescribing ezetimibe at this time are very persuasive, especially with regard to the cost of preventing cardiovascular events with this drug, even if the recent reports of such efficacy are eventually verified. This kind of perspective from a cardiovascular scientist is most beneficial to general internists like me.

Your comments regarding rosiglitazone form a sort of inverse analogy to ezetimibe; if physicians who prescribed rosiglitazone are not to be condemned for "guessing wrong" in advance of the data proving that drug was harmful, then physicians who prescribed ezetimibe should not be congratulated for "guessing right" if this drug is proven to be beneficial. Good point.

However, I chose never to prescribe rosiglitazone, even before its harmful outcomes were known. My decision was not due to luck or clairvoyance, but to the simple fact that rosiglitazone raises LDL cholesterol, and a very similar alternative drug (pioglitazone) exists which lowers LDL cholesterol instead. Knowing that every point of LDL increase in a diabetic is correlated with increased cardiovascular risk, I thought it would be folly to choose an agent which worsened LDL, even if it was only a surrogate marker.

So, all of these examples get back to the question of the strength of LDL-lowering as a surrogate marker for reduction of cardiovascular events. The example of rosiglitazone versus pioglitazone seems to strengthen LDL as a valid surrogate marker for events. It appears that the new data for ezetimibe will also do so, if it confirms that this weak LDL-lowering drug also weakly improves cardiovascular event rates. Of course, ezetimibe still may not be worth its cost, as you point out, and that appears to be the bottom line, at least until ezetimibe is available as a cheap generic medication.

On 2015 Feb 03, Todd Lee commented:

Dr. Keller,

You do make a very interesting point about hindsight.

It is a fact that billions of dollars have been spent in North America on a drug that, to date, does not have any high quality published evidence of efficacy in terms of clinical endpoints. This drug was thus used on the basis of theoretical arguments and a belief that lower LDL means lower events. The findings of IMPROVE-IT may suggest that some patients that have received this drug all of this time did derive benefit. However, based on the effect size in the trial and the patients enrolled, this may be the minority -- and at a considerable health care cost for each benefit.

I'm really not sure what to infer about physicians who believed in the drug all this time. Are they to be commended for being correct?

I'm not certain that money could not have been used for greater benefit with another intervention -- that is a question for health care economists; however, at $1,000,000 per event prevented (earlier post) it seems likely there may have been a more cost effective option.

Let me contrast that with another example as I believe the case for ezetimibe and LDL is analogous to what people believed about hemoglobin A1c and the complications of diabetes. ROSIGLITAZONE was an extremely popular drug in 2006 -- to the tune of US prescriptions exceeding 2 billion dollars. This was because people inferred that lower A1c meant improved outcomes. Later, after the landmark paper showed that ROSIGLITAZONE might have been associated with increased cardiovascular risks, the sales plummeted to less than 15 million dollars in 2012. Thousands of lawsuits were settled and billion dollar fines were paid (for ROSIGLITAZONE and issues around other drugs) by the manufacturer. Furthermore, that drug was withdrawn from numerous European markets.

I'm not sure what to infer about all of the physicians who believed in that drug. Are they all to be condemned? The answer is no. Physicians do the best they can for their patients within the limits of current medical knowledge. They had no a priori way of knowing that this intervention may actually harm patients more than help them.

That said, it is not an accident that both agents were billion dollar winners for their prospective companies and that they were heavily marketed.

Our group incidentally received a marketing pamphlet on ezetimibe in the mail today extolling the IMPROVE-IT trial. The jist of it: the lipid hypothesis is reaffirmed and ezetimibe has been proven effective in patients at risk of coronary artery disease. Does it matter that the paper hasn't been peer reviewed or put in the context of all of the negative papers before it? Not at all. That wouldn't be very good marketing.

On 2015 Jan 24, David Keller commented:

Thank you, Dr. Lee, for that thorough, informative and excellent reply to my comment. For many years, I avoided prescribing ezetimibe (either alone or with a statin) due to the lack of data documenting improved outcomes. Prior to IMPROVE-IT, I would not have questioned the need for your study, and I would have been curious to learn why other clinicians were prescribing ezetimibe despite its interesting ability to improve lipids but not outcomes. Other non-statins which improve lipids - such as niacin, fibrates and bile acid sequestrants - had been able to demonstrate improved cardiovascular outcomes, at least in certain populations or under certain circumstances. The reports of improved outcomes from IMPROVE-IT seemed to reaffirm the importance of lowering LDL and to remove the major obstacle which had prevented me from adding ezetimibe when patients failed to achieve their LDL goal on a tolerable dose of statin. Based on the reservations you have expressed regarding IMPROVE-IT, I will hold back from prescribing it until the questions you raised are answered.

For the sake of discussion, assume that cardiovascular outcome benefits of ezetimibe are eventually proved to your satisfaction. How, then, would that affect the utility of your investigation of why physicians were prescribing a drug which lacked outcome data, if their patients were thereby benefiting from lower event rates? These physicians may have prescribed ezetimibe based on their anecdotal experience with the drug, their knowledge of lipid physiology and pharmacology, or other factors, making them guilty of nothing more than being ahead of the randomized trial data, correctly predicting the outcome benefits of ezetimibe, and benefiting their patients with better outcomes sooner than more conservative physicians like me. Should we be trying to "correct" that kind of "mistake"?

On 2015 Jan 24, Todd Lee commented:

As stated in our paper, we need the data from IMPROVE-IT to better inform us on the use of ezetimibe. However, rather than relying on a conference presentation and press releases we will also need to await the peer-reviewed and sponsor-independent analysis of the IMPROVE-IT trial to judge the quality of the results and evaluate their impact on general practice. Given interim analysis was performed (and the study was subsequently re-sized) any multiple comparisons performed will require expert statistical review.

Furthermore, given no other positive studies exist, it may be prudent to perform an independent individual patient data analysis of all similar studies to better refine or confirm the estimate of effect prior to making any final conclusions from one trial.

The net conclusion of this study, if the presented data is taken at face value, is a number needed to treat (NNT) of 50 over 7 years for the composite outcome. Overall mortality was not reduced. At generic Canadian prices it would cost approximately $58,765 to prevent 1 event over 7 years (Ontario formulary price as of January 2015). However, at US brand name prices of approximately $8.50 per day (Lexicomp 2015) the cost of preventing one composite outcome would be more than $1,000,000.

It is also important to note that IMPROVE-IT was a secondary prevention study (acute event within 10 days) and not primary prevention. In primary prevention, the NNT is likely much higher and the corresponding costs per event prevented would increase proportionately and likely be substantial even at generic prices. In our cohort 6/17 (35%) were receiving ezetimibe for primary prevention.

Whether the drug lowers event rates in the absence of a statin remains unproven and cannot be inferred from this study. Nonetheless, it will be interesting to see the effects on monotherapy uptake given the publicity around this study and also when IMPROVE-IT is ultimately published.

The impact of this study on the uptake of other drugs approved on the basis of LDL as a surrogate marker is also not to be underestimated. The issue of treating to specific LDL targets is currently being debated amongst experts after recent changes to the guidelines. It would be somewhat naive to think that there isn't a substantial market pressure behind bringing back targets to be measured and obtained through additional medications.

On 2015 Jan 23, David Keller commented:

Recent evidence of ezetimibe outcome benefits undermines the premise of this study

The underlying premise of this study of ezetimibe prescribing patterns is that there is “a lack of evidence supporting its efficacy” in reducing adverse cardiovascular outcomes, despite its demonstrated efficacy in improving lipids and other surrogate endpoints (1).

In the recently reported “IMPROVE-IT” study, patients taking ezetimibe plus simvastatin experienced significantly fewer major cardiovascular events (as measured by a composite of cardiovascular death, non-fatal myocardial infarction, non-fatal stroke, rehospitalization for unstable angina or coronary revascularization occurring at least 30 days after randomization) than patients treated with simvastatin alone (2).

Now that there is evidence that ezetimibe improves cardiovascular outcomes, the question of why clinicians prescribed it inappropriately in the past seems moot. Perhaps their clinical experiences using ezetimibe convinced them that evidence of beneficial outcomes would emerge over time. If ezetimibe was improving outcomes all along, then studies based on the premise that it should not have been used seem less relevant.

References:

1: McDonald EG, Saleh RR, Lee TC. Ezetimibe use remains common amongst medical inpatients. Am J Med. 2014 Oct 19. pii: S0002-9343(14)00917-6. doi: 10.1016/j.amjmed.2014.10.016. [Epub ahead of print] PubMed PMID: 25448168.

2: Kohno T. Report of the American Heart Association (AHA) Scientific Sessions 2014, Chicago. Circ J. 2014 Dec 15. [Epub ahead of print] PubMed PMID: 25502168.

On 2017 May 21, Misha Koksharov commented:

This paper provides interesting insights on the role of the 471–481 and 487–495 flexible regions in P. pyralis luciferase and on their utility in improving its thermostability.

For those interested to investigate this topic further, I suggest to look also at the substitution D475P. As can be seen in the alignments below, this substitution occurs naturally in Luciola and in many other beetle luciferases:

D476: http://forumbgz.ru/user/upload/file33577.gif

D489: http://forumbgz.ru/user/upload/file33578.gif

Moreover, in the available structures of L. cruciata luciferase the region 471–481 is not flexible - in contrast to P. pyralis enzyme - so I expect that this mutation should work (where the mutations D476P (Yu H, 2015) and D476N (Amini-Bayat Z, 2012) were inefficient). Interestingly, the successful substitution H489P also naturally occurs in many pH-insensitive luciferases (as shown in the alignments above).

On 2015 May 26, Paul Brookes commented:

Portions of Figure 4 and Figure 5 from this paper appear more similar than would be expected by chance, when compared to portions of Figure 3 and Figure 8, from Zhao ZQ, 2012 Br J Pharmacol. 167, 1550-1562, PMID: 22823335

Relevant information is available in the following two images... http://imgur.com/0DI1G78 http://imgur.com/u47Fdr8

Usual disclaimers apply - i.e. go look for yourself at the originals and decide if they're similar, don't take my word for it, no implications about motives or underlying causes of this apparent similarity should be taken from this comment.

That being said, the lead author has had two papers retracted for similar matters... http://retractionwatch.com/2015/05/26/heart-repair-study-retraction-marks-second-for-mercer-unviersity-researcher/

On 2016 Aug 24, Ben Goldacre commented:

This trial has the wrong trial registry ID associated with it on PubMed: both in the XML on PubMed, and in the originating journal article. The ID given is NCT097115. We believe the correct ID, which we have found by hand searching, is NCT01857388.

This comment is being posted as part of the OpenTrials.net project^[1] , an open database threading together all publicly accessible documents and data on each trial, globally. In the course of creating the database, and matching documents and data sources about trials from different locations, we have identified various anomalies in datasets such as PubMed, and in published papers. Alongside documenting the prevalence of problems, we are also attempting to correct these errors and anomalies wherever possible, by feeding back to the originators. We have corrected this data in the OpenTrials.net database; we hope that this trial’s text and metadata can also be corrected at source, in PubMed and in the accompanying paper.

Many thanks,

Jessica Fleminger, Ben Goldacre*

[1] Goldacre, B., Gray, J., 2016. OpenTrials: towards a collaborative open database of all available information on all clinical trials. Trials 17. doi:10.1186/s13063-016-1290-8 PMID: 27056367

* Dr Ben Goldacre BA MA MSc MBBS MRCPsych<br> Senior Clinical Research Fellow<br> ben.goldacre@phc.ox.ac.uk<br> www.ebmDataLab.net<br> Centre for Evidence Based Medicine<br> Department of Primary Care Health Sciences<br> University of Oxford<br> Radcliffe Observatory Quarter<br> Woodstock Road<br> Oxford OX2 6GG

On 2015 Jul 24, Md. Shahidul Islam commented:

The 2015 Islet Society Meeting took place in Sydney during 19-21 July, 2015. This meeting, organised by Anand Hardikar was also of high scientific quality. The 2016 Islet Society meeting will be in Istanbul (15-17 July, 2016). The organizer of 2016 meeting is Erdal Karaoz

On 2015 Mar 05, Anders Prior commented:

High effect sizes and immortal person-time

In a paper published in Medicine, Wang et al. presented this cohort study of 47,225 stroke patients.¹ (Wang JY, 2014) They concluded that the 30-day mortality after stroke was significantly reduced if treatment with antipsychotic drugs was initiated before or after the stroke.

The reported effect sizes were very large; the mortality rate in stroke patients decreased by 73% in those who received antipsychotic treatment before the stroke and by 87% in those who received antipsychotic treatment shortly after the stroke. Effects of this size are rarely seen in such studies.

The study was based on a nested case-control design. Included cases were stroke patients who died within 30 days after having a stroke, while controls (matched on age, gender and stroke date) were stroke patients who survived at least 30 days after their stroke. The main exposure was antipsychotic drugs given at any time within 30 days after the stroke date. Multivariate logistic regression was used to calculate odds ratios of mortality.

We are concerned that the results may be substantially biased because the definition of the antipsychotic user group conditions on the future; in order to receive the antipsychotic treatment and become a member of the antipsychotic user group, study participants need to survive until the drug is prescribed. In other words, study participants in the antipsychotic user group are ‘immortal’ until the day of treatment (immortal person-time).² The authors do not describe any analytical measures taken to counteract this conditioning. If they have not taken this into account, this would pose a very serious problem as also described numerous times in the epidemiological literature.³ (Hanley JA, 2014) Immortal time bias will generate an illusion of treatment effectiveness and is frequently found in observational studies that compare with non-users.⁴ (Suissa S, 2007)

In general, suspicion for immortal time bias should be raised when exposure groups are assigned with no regard to exposure time in a longitudinal study. Furthermore, the reported effect sizes are surprisingly high, especially when considering the fragility of the population in question. This group consists of stroke patients with complications; they may suffer from e.g. post-stroke delirium and may need antipsychotic treatment. They would most likely have more adverse outcomes, not the opposite.^5,6 (Shi Q, 2012, Prior A, 2014)

Dr. Anders Prior, MD

Research Unit for General Practice and Section for General Medical Practice, Department of Public Health, Aarhus University, Denmark

Dr. Thomas Munk Laursen, PhD

National Centre for Register-based Research, Department of Economics and Business, Aarhus University, Denmark

Prof. Mogens Vestergaard, PhD

Research Unit for General Practice and Section for General Medical Practice, Department of Public Health, Aarhus University, Denmark

References

1 Wang JY, Wang CY, Tan CH, Chao TT, Huang YS, Lee CC. Effect of different antipsychotic drugs on short-term mortality in stroke patients. Medicine (Baltimore). 2014;93(25):e170.

2 Rothman KJ, Greenland S, Lash TL. Modern Epidemiology. 3rd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2008.

3 Hanley JA, Foster BJ. Avoiding blunders involving 'immortal time'. Int J Epidemiol. 2014;43(3):949-961.

4 Suissa S. Immortal time bias in observational studies of drug effects. Pharmacoepidemiol Drug Saf. 2007;16(3):241-249.

5 Shi Q, Presutti R, Selchen D, Saposnik G. Delirium in acute stroke: A systematic review and meta-analysis. Stroke. 2012;43(3):645-649.

6 Prior A, Laursen TM, Larsen KK, et al. Post-stroke mortality, stroke severity, and preadmission antipsychotic medicine use--a population-based cohort study. PLoS One. 2014;9(1):e84103.

On 2015 Jan 12, Larry Parnell commented:

This paper was my selection for Paper of the Week for 1-5 Dec 2014.

This is highly important work and the conclusions drawn are valuable and applicable to many areas of genetic research into complex phenotypes. It is great to see scientists of this stature adding important insight into the role of GxEs, especially in terms of influence on gene expression.

I don't feel that the relationship described in this paper for ADIPOQ is so meaningful as the variant described is not in high LD in most populations with the ADIPOQ variants that support known GxE interactions for cardiometabolic phenotypes. See http://www.biodatamining.org/content/7/1/21 Nonetheless, there is also in the BioData Mining article a list of GxEs described for UCP2, which does have evidence in this report above.

On 2016 Aug 23, Ben Goldacre commented:

This trial has the wrong trial registry ID associated with it on PubMed: both in the XML on PubMed, and in the originating journal article. The ID given is NCT0093283. We believe the correct ID, which we have found by hand searching, is NCT00932893.

This comment is being posted as part of the OpenTrials.net project^[1] , an open database threading together all publicly accessible documents and data on each trial, globally. In the course of creating the database, and matching documents and data sources about trials from different locations, we have identified various anomalies in datasets such as PubMed, and in published papers. Alongside documenting the prevalence of problems, we are also attempting to correct these errors and anomalies wherever possible, by feeding back to the originators. We have corrected this data in the OpenTrials.net database; we hope that this trial’s text and metadata can also be corrected at source, in PubMed and in the accompanying paper.

Many thanks,

Jessica Fleminger, Ben Goldacre*

[1] Goldacre, B., Gray, J., 2016. OpenTrials: towards a collaborative open database of all available information on all clinical trials. Trials 17. doi:10.1186/s13063-016-1290-8 PMID: 27056367

* Dr Ben Goldacre BA MA MSc MBBS MRCPsych<br> Senior Clinical Research Fellow<br> ben.goldacre@phc.ox.ac.uk<br> www.ebmDataLab.net<br> Centre for Evidence Based Medicine<br> Department of Primary Care Health Sciences<br> University of Oxford<br> Radcliffe Observatory Quarter<br> Woodstock Road<br> Oxford OX2 6GG

On 2015 Jan 07, Mohammad Salman commented:

I'm surprised by the findings of this paper since several authors have reported antiinflammatory effects of Thymoquinone and inhibition of COX-2 and PGE2 via Thymoquinone. See:

1: Kundu JK, Liu L, Shin JW, Surh YJ. Thymoquinone inhibits phorbol ester-induced activation of NF-κB and expression of COX-2, and induces expression of cytoprotective enzymes in mouse skin in vivo. Biochem Biophys Res Commun. 2013 Sep 6;438(4):721-7. doi: 10.1016/j.bbrc.2013.07.110. Epub 2013 Aug 1. PubMed PMID: 23911786.

2: Al Wafai RJ. Nigella sativa and thymoquinone suppress cyclooxygenase-2 and oxidative stress in pancreatic tissue of streptozotocin-induced diabetic rats. Pancreas. 2013 Jul;42(5):841-9. doi: 10.1097/MPA.0b013e318279ac1c. PubMed PMID: 23429494.

3: El Mezayen R, El Gazzar M, Nicolls MR, Marecki JC, Dreskin SC, Nomiyama H. Effect of thymoquinone on cyclooxygenase expression and prostaglandin production in a mouse model of allergic airway inflammation. Immunol Lett. 2006 Jul 15;106(1):72-81. Epub 2006 May 22. PubMed PMID: 16762422.

4: Marsik P, Kokoska L, Landa P, Nepovim A, Soudek P, Vanek T. In vitro inhibitory effects of thymol and quinones of Nigella sativa seeds on cyclooxygenase-1- and -2-catalyzed prostaglandin E2 biosyntheses. Planta Med. 2005 Aug;71(8):739-42. PubMed PMID: 16142638.

On 2014 Dec 06, Yannis Karamanos commented:

This was not a "happy" idea to use ENGase for endoglucanase... This abreviation is in use, for many years, for endo-N-acetyl glucosaminidase

On 2015 Aug 17, Anders von Heijne commented:

It is so interesting to compare the workings of our different healthcare systems. The type of order/referral for overread of outside radiological stuies described in this paper have been used in swedish radiology for at least as long as I have been in practice - ie at least for three decades. In addition to the obvious advantages described by the authors, the clinican that places the overread order can add relevant facts and queries in the new order. One easy way of documenting the outside report is to add it as a DICOM image to the study and save it in the PACS, rather than transfering it between different RIS.

On 2016 Feb 18, Kristina Hanspers commented:

The pathway in figure 2 is available in the "Open Access Publication" collection at WikiPathways: http://www.wikipathways.org/index.php/Pathway:WP3297. This pathway can be downloaded for use in network analysis tools such as Cytoscape and PathVisio.

On 2015 Oct 22, Horacio Rivera commented:

A de novo inv dup(1) turns out to be a rea(1)dup q chromosome I remark that the de novo mosaic 1q32→qter duplication onto 1pter (concomitant with a normal clone) described by Levy et al. [2015] is not an inverted duplication because direct and inverted duplications are officially defined as “a gain of a chromosomal segment observed at the original chromosome location” [Shaffer et al., 2013], not to mention that the orientation of the extra segment was indeed a direct one. It is significant that in their discussion, Lévy et al. [2015] refer to three other similar chromosome-1 rearrangements entailing an 1q duplication (although none with interstitial telomeric repeats) and visualize them as “recombinants” from an hypothetical pericentric inversion. In this regard, a compilation of 104 recombinant-like chromosomes of de novo or sporadic occurrence [Rivera et al., 2013] lists 11 other comparable chromosome-1 composites entailing dup q/del p but without an interstitial telomere. To avoid a nonsensical “der vs rec” controversy, we have designated such rearranged chromosomes with the official term rea coupled with the lengthy description of the novel composite [Rivera et al., 2013]. In that paper, we pointed out that “this formula makes no causal assumptions, unambiguously describes the rearranged chromosome, allows for a meiotic or mitotic origin, and is consistent with the involvement of 1 or 2 homologs”. Moreover, the term rea had already been used for this purpose [Thomas et al., 2006] and properly describes a rearranged unbalanced chromosome mimicking a recombinant ensued from a pericentric inversion as it is epitomized by the rea(1)(qter→q32::pter→qter) here alluded to. It goes without saying that inv dup is also improperly used to designate mirror structures such as isodicentrics and neocentric isofragments [e.g., Warburton et al., 2000]. Although Lévy et al. [2015] recognized that “[T]he recurrent nature of all these similar recombinants, including our dup(1q), suggests an identical mechanism of formation”, they failed to identify it. According to D’Angelo et al. [2009], who analyzed in fine detail two dup q/del p and two other comparable chromosome-1 rearrangements, the DNA repair mechanism of non-homologous end joining (NHEJ) appears to be “the pathway in the formation of these de novo nonreciprocal translocations, because of the lack of evidence to support a homology-based recombination mechanism”. Yet, the location in unique, non-repetitive DNA sequences of all the breakpoints in the four chromosome-1 rearrangements above mentioned [D’Angelo et al., 2009] may call into question the NHEJ mechanism for rearranged chromosomes with an interstitial telomere alike to the exceptional rea(1) documented by Lévy et al. [2015]. Because Lévy et al. [2015] also omitted some relevant references on other rearrangements with interstitial telomeres, I reiterate here the academic and moral duty that authors, reviewers, editors, and readers have to improve the current citation practices [Rivera, 2014]. Finally, I stress that seven months ago a Letter to the Editor with these comments was judged unacceptable by the concerned journal because “the conclusions of Dr. Levy's paper are really not about terminology and nomenclature”. REFERENCES D’Angelo CS, Gajecka M, Kim CA, Gentles AJ, Glotzbach CD, Shaffer LG, Koiffmann CP. 2009. Further delineation of nonhomologous-based recombination and evidence for subtelomeric segmental duplications in 1p36 rearrangements. Hum Genet 125:551-563. Lévy J, Receveur A, Jedraszak G, Chantot-Bastaraud S, Renaldo F, Gondry J, Andrieux J, Copin H, Siffroi J-P, Portnoï M-F. 2015. Involvement of interstitial telomeric sequences in two new cases of mosaicism for autosomal structural rearrangements. Am J Med Genet Part A 167A:428-433. Rivera H. Commentary: peer review and incomplete reference lists. 2014. Account Res 21:138-141. Rivera H, Domínguez MG, Vásquez-Velásquez AI, Lurie IW. 2013. De novo dup p/del q or dup q/del p rearranged chromosomes: review of 104 cases of a distinct chromosomal mutation. Cytogenet Genome Res 141: 58-63. Shaffer LG, McGowan-Jordan J, Schmid M. 2013. ISCN 2013: an international system for human cytogenetic nomenclature (2013), Basel, S Karger. p. 69. Thomas NS, Durkie M, Van Zyl B, Sanford R, Potts G, Youings S, Dennis N, Jacobs P. 2006. Parental and chromosomal origin of unbalanced de novo structural chromosome abnormalities in man. Hum Genet 119: 444-450. Warburton PE, Dolled M, Mahmood R, Alonso A, Li S, Naritomi K, Tohma T, Nagai T, Hasegawa T, Ohashi H, Govaerts LC, Eussen BH, Van Hemel JO, Lozzio C, Schwartz S, Dowhanick-Morrissette JJ, Spinner NB, Rivera H, Crolla JA, Yu C, Warburton D. 2000. Molecular cytogenetic analysis of eight inversion duplications of human chromosome 13q that each contain a neocentromere. Am J Hum Genet 66:1794-1806.

On 2015 Jan 26, Vojtech Huser commented:

This is a report on a clinical trial and it is mentioned in the abstract (NCT02231866). Ideally, the identifier would be within the PubMed [SI] field. This results in this result article not displaying as linked on ClinicalTrials.gov

On 2014 Nov 27, Jürgen Hänggi commented:

Dear community

this new study shows that a recently published "apparent" sex effect in the form of "increased interhemispheric connectivity in women and increased intrahemispheric connectivity in men" (see http://www.ncbi.nlm.nih.gov/pubmed/24297904) is driven by brain size and not by sex per se.

On 2015 Sep 03, Lydia Maniatis commented:

It seems generally to be the case that when a darkish-looking surface grows in area, it appears to lighten. There is no doubt that this occurs, and also no doubt that the effect is subject to a large degree of variability between experiments and between individuals. Boyaci et al (2014) report, as their main finding, that the effect occurs in the context of "computer-rendered scenes" (their stimuli) as well as in the context of "real scenes" (referring to the rarefied laboratory set-up of Radonjic and Gilchrist (2014)).

As a way of describing experimental conditions - whose results, moreover, were subjected to technical analyses to determine which datasets fit a linear model, a quadratic model, a one-phase model, a two-phase model, and used to form hypotheses about the neurophysiological underpinnings of performance and to judge conceptual models - the crude distinction between "computer-rendered" and "real" seems rather insufficient. Indeed, when it comes anything more specific than the general consistency with the "larger gets lighter" effect, all bets are off. Noting that their results were different from those of Radonjic and Gilchrist (2014), the authors speculate simply that: "The disagreement between our studies is likely to be because of differences between the stimuli." More specifically: "Whereas their stimuli (Radonji砦 Gilchrist, 2014) were real and as simple as possible, ours were computer generated and relatively more complex." Like the unqualifed "real" vs "computer-generated" distinction, the "simple vs complex" distinction is too vague for the purpose. Very subtle and neurophysically complex perceptual operations can be triggered by geometrically simple stimuli. Surely, mathematical modelling and theoretical/neurophysiological extrapolations should follow, not precede, qualitative understanding of a phenomenon and the effects of conditions. Otherwise, such speculation is strictly ad hoc.

The "area rule" as originally proposed said more than that the darker of two areas will lighten with area. It said that this lightening of the darker area will "anchor" cause the appearance of the lighter area such that it brightens and eventually becomes luminous. The proposal was made in order to rationalize the luminosity observed in the disc in classic disc/annulus experiments, without having to ascribe a role to figure-ground organization. The predicted changes in the lightness of the lighter area qua area apparently haven't been corroborated and are no longer referred to or tested for. As an isolated phenomenon not affecting surrounding surfaces, it's hard to see the theoretical importance of a highly-variable and condition-sensitive tendency toward lightening of relatively darker surface with increases in area, or the value of making fussy models of this tendency, tailored to particular stimuli.

I would also like to take issue with the choice of references the authors chose to support their opening assertion that "The lightness of a surface depends not only on its luminance but also on its geometry and the context within which it is viewed (Boyaci, Doerschner, Snyder, & Maloney, 2006; Gilchrist, 2006; Kingdom, 2011; Maloney, Boyaci, & Doerschner, 2005; Maloney, Gerhard, Boyaci, & Doerschner, 2010)." The case for this was made in classic literature and experiments well before 2006. The assertion is so fundamental and uncontroversial that references are not even necessary.

On 2015 Nov 04, Lydia Maniatis commented:

Continuation of "area rule" story: Avoiding acknowledging the role of figure-ground in contrast effects also allowed advocates of "anchoring theory" to claim that "the debate seems to revolve around layer models and framework models" (Gilchrist, 2015), frameworks being adjacent "like countries on a map." Figure-ground structure means layers, and so would complicate this presentation.

On 2015 Sep 20, Lydia Maniatis commented:

In line with the earlier point about avoiding figure/ground issues, the "multi-sector" stimuli used here are of the sort used by Gestalt psychologists to demonstrate and study figure-ground effects (e.g. the role of sector color, size, orientation; bistability). But such effects and their potential lightness consequences are not acknowledged.

On 2015 Sep 17, Lydia Maniatis commented:

"Simple" vs "complex" stimulus is not a meaningful distinction, because there is no objective criterion for making it. Geometric simplicity may or may not produce complexity in the percept, of different kinds.

Which is more simple: a. the three pacmen of a Kanizsa triangle, that produce subjective contours/ contrast/amodal completion; b. the simultaneous contrast effect that produces contrast and amodal completion? c. the de Valois checkerboard that produces assimilation; d. a random array of variously shaped, overlapping white, grey, black shapes? e. the checkerboard used here, which produces transparency/inhomogenous illumination effects; f.a black shape that produces the impression of two overlapping black shapes; g. etc.

The authors offer their own “clear distinction” of simple vs complex. Not only is this proposed distinction unclear, it lacks theoretical content. The authors subsequently seem to ignore it, and use the terms in a loose and undefined way. The proposed distinction, attributed to “anchoring theory” is: “A simple image is one in which all the surfaces lie within a single illumination level (a single framework) whereas complex images contain multiple adjacent fields of illumination (multiple frameworks).” This distinction is theoretically hollow, because it references actual (actual level of illumination) and not perceived (perceived illumination) image features. By this definition, a photograph of sunlight and shadow should count as a single framework, provided it is being viewed under a homogeneous illumination. Even if we take the authors' distinction to refer to perceived illumination (which begs the question the “frameworks” argument is supposed to answer, i.e. how do we parse the scene into separate illumination “frameworks”), they don't stick to it. A little later, for example, we are told that the stimuli of Boyaci et al (2014) “are still fairly simple, [but] qualify [] as complex as they consist of more than one framework” (the distinction is invoked to explain discrepancies between those authors findings and the findings of R and G (2014). But Boyaci et al's stimuli are neither under inhomogeneous illumination nor are they designed to create such an impression. So the definition of “framework,” which wasn't explanatory to begin with, has shifted in the space of a few paragraphs, from “framework of illumination” to something else (what?). We are similarly told that “a dozen previous experiments...used images that qualify as complex images;” but many (perhaps all) of these studies did not involve inhomogeneous illumination, real or perceived.

On 2015 Sep 17, Lydia Maniatis commented:

A short, instructive history of the “area rule”

The “area rule” was born in an attempt to deny the role of figure-ground structure in lightness. This is its original sin, and has led to interesting distortions in theory and practice.

Classic disc-annulus experiments, which demonstrated the dependence of lightness on luminance ratios, also showed an asymmetrical influence of disc and annulus. The annulus would always look white, and push a lighter disc towards luminosity. In other words, raising the luminance of the annulus would lighten the disc, but not vice versa. The disc appeared to lie on top of an amodally-completed larger disc, so an easy provisional conclusion would be to assign different influences to figure and ground in mediating lightness perception of a surface (the same asymmetry applies to figure-ground contrast in general).

Gilchrist et al (1999) did not like this solution because it interfered with their preferred theoretical assumption that the highest luminance in a (vaguely-defined) “framework” would be white. Clearly, in the very simple, disc-annulus situation, the highest luminance was not necessarily white. Instead of acknowledging a role for figure-ground, Gilchrist et al (1999) created a new rule, stating that if the darker area was more than 50% larger than the smaller, then it would lighten, and progressively push the smaller, lighter area toward luminosity. They presented a speculative function, reprinted in Gilchrist and Radonjic (2009), that has never been corroborated, despite a number of attempts.

The Gilchrist group's own results constantly cried out for a figure-ground explanation. Tellingly, they were forced to modify their area claim to include “amodally-completed” area – thus in effect making the area rule indistinguishable from a figure-ground claim. Later, Economou et al (2007) acknowledged a similar, figure-ground-related asymmetry in the simultaneous contrast display. The team acknowledged the asymmetry but did not explore it further.

Preserving the highest-luminance-white rule was not the only or even most important incentive for rejecting a possible figure-ground role. Another fundamental claim of Gilchrist et al (1999) was that the classic simultaneous contrast demonstration is due to a process which, at a certain stage, treats each square and its interior as a separate “framework” and evaluates its contents based on the ratio principle and highest-luminance rule. The idea that the lightness of the targets is actually mediated by local luminance contrast between the apparent figure and its background was not compatible with this assumption. However, it is easy to show (Maniatis, 2015), by adding surfaces within each putative “framework” that border contrast between figure and ground, not the ratios with all surfaces contained in the background square, mediates this effect.

The commitment to avoiding acknowledging a role for figure-ground explains, I believe, the preference manifested by investigators with these theoretical commitments for stimuli which either did not produce figure-ground effects, or in which the contrast effects would average out. Specifically, they adopted the use of checkerboards or Mondrians, and random or semi-random selection of luminances. Such stimuli and choices muddy rather than clarify the role of structure in lightness. Thus, proponents of a “Gestalt” theory, were, paradoxically forced by their commitments to prefer stimuli in which image structuring could be ignored.

There was a second reason that this “Gestalt” theory needed to avoid confronting the role of structure in lightness, and this was that it did not/could not address the fact that we sometimes perceive surfaces as lying beneath transparent layers with their own lightness. As in the case of figure-ground/amodal completion effects, such layers arise when contours showing good continuation intersect, with the added proviso that the luminance structure is compatible with such a solution. Checkerboards, lacking such cues, avoid such effects.

Well, actually, they don't. They often produce multiple such effects, as well as luminosity. This latter result arose in Radonjic et al (2011). It was awkward and they tried to explain it away by selectively attributing inconvenient results to presentation on an “emissive” screen. Allred et al (2012) tentatively acknowledged, after much highly technical wrestling with very low-resolution data, the self-evident yet apparently unplanned-for fact that checkerboards do produce differential lightness impressions. So restricting the class of allowable stimuli (rationalized on the basis that they were “simple” and that results would carry over to more “complex” situations – a view codified in the oft-repeated “applicability assumption”) in order to avoid confronting figure-ground and transparency effects has nevertheless led researchers back to these same, unavoidable issues.

It is interesting that the checks on a checkerboard can coalesce into transparent overlays/underlying surfaces despite the absence of apparent overlap. It is surely not unrelated to the fact that checkerboards produce assimilation rather than contrast when we replace a black or white check with a grey one (the de Valois and de Valois checkerboard contrast demonstration). It would be worth analyzing.

On 2015 Aug 12, Lydia Maniatis commented:

This article is a bit misleading. The authors seem to be claiming to have corroborated the area rule, at least for the case where the darker area fills more than half the visual field.

As defined and illustrated by Gilchrist et al (1999) and Gilchrist and Radonjic (2009), the area rule predicts the onset of luminosity when the darker area exceeds half the display. But this prediction has never been corroborated, even for far, far larger darker area coverage, nor has the description of the rule changed. In addition, lightening of the darker are has never been shown to start until the latter covers far more than 50% of total area. So the claim that this study has corroborated the "area rule" needs to be qualified by the introduction of a new description of the area rule.

On 2013 Oct 25, DAVID SANDERS commented:

From the abstract of Ueda T, 1989

"Two proteins stimulating the GTPase activity of the smg-21 GTP-binding protein (smg p21) having the same effector domain as the ras proteins (ras p21s) are partially purified from the cytosol fraction of human platelets. These proteins, designated as smg p21 GTPase activating protein (GAP) 1 and 2, do not stimulate the GTPase activity of c-Ha-ras p21. smg p21 GAP1 and 2 are separated from c-Ha-ras p21 GAP by column chromatographies. The activity of smg p21 GAP1 and 2 is killed by tryptic digestion or heat boiling. The Mr values of smg p21 GAP1 and 2 are similar and are estimated to be 2.5-3.5 x 10(5) by gel filtration analysis. These results indicate that there are two GAPs for smg p21 in addition to a GAP for c-Ha-ras p21 in human platelets."

On 2014 Dec 10, Alberto Zambrano commented:

Dear Ilya, The Ab used was initially described by Dr. Sonenberg (Mol Cell Biol. 1998 Jan;18(1):334-42, A novel functional human eukaryotic translation initiation factor 4G.) and was kindly donated by Dr. Cesar de Haro (CBM, Madrid, Spain) There you can see that the size of the recombinant form of EIF4G2 (His-tagged) (Figure 3) recognized by the Ab is a little bit lower than 206 KDa. You can also see another prominent band higher than 117KDa, recognized by the Ab. In a another article from the same author (J Neurosci. 2012 Apr 18;32(16):5620-30. doi: 10.1523/JNEUROSCI.0030-12.2012. Regulation of neuronal mRNA translation by CaM-kinase I phosphorylation of eIF4GII) the same Ab was also used but, unfortunately the protein markers were not indicated. In our blots, the Ab reactivity shows either a prominent, discrete band or a doublet at size indicated. We don't observe any other significant reactivity in the full blots. In addition, those bands were sensitive to specific specific siRNA downregulation (figure 5). We don't have a clue about the origin of such differential electroforetic mobility differencies, described in the original paper or in ours (in MEFs, with respect to the predicted size but with the evidences we had, we assumed that protein identity.

On 2014 Dec 01, Ilya M Terenin commented:

Ghm... Guys, why your eIF4G2 is heavier than 150 kDa on blots? The protein is ~100 kDa.

On 2014 Dec 03, Rafael Najmanovich commented:

We predicted back in 2012 (Chartier M, 2012) as part of a large scale analysis of rare codon cluster that such clusters may play a role in the molecular recognition of the nascent protein for intracellular targeting and membrane insertion. It is unfortunate that the authors were not aware of our publication.

On 2015 Jan 03, William Grant commented:

Here are some papers not included in this review showing benefits of vitamin D testing. Peiris AN, Bailey BA, Grant WB, Mascitelli L. Vitamin D testing. Lancet 2012 May 5;379:1699-1701.

Der T, Bailey BA, Youssef D, Manning T, Grant WB, Peiris AN.,Vitamin D and prostate cancer survival in veterans. Military Med. 2014;179 (1) :81–84.

Peiris AN, Bailey BA, Manning T. Relationship of vitamin D monitoring and status to bladder cancer survival in veterans. South Med J. 2013 Feb;106(2):126-30.

Bailey BA, Manning T, Peiris AN. Vitamin D testing patterns among six Veterans Medical Centers in the Southeastern United States: links with medical costs. Mil Med. 2012 Jan;177(1):70-6.

Of course some may have been published after your literature search was completed. However, they do support the benefits of vitamin D testing among hospital patients.

Also, this paper is supportive in that it found that patients in the ICU with very low 25OHD concentrations benefited from vitamin D supplementation. Amrein K, Schnedl C, Holl A, Riedl R, Christopher KB, Pachler C, Urbanic Purkart T, Waltensdorfer A, Münch A, Warnkross H, Stojakovic T, Bisping E, Toller W, Smolle KH, Berghold A, Pieber TR, Dobnig H. Effect of High-Dose Vitamin D3 on Hospital Length of Stay in Critically Ill Patients With Vitamin D Deficiency: The VITdAL-ICU Randomized Clinical Trial. JAMA. 2014 Oct 15;312(15):1520-30.

One of the impediments to acceptance of vitamin D seems to be the lack of supportive trials. The trials have not supported the ecological and observational studies largely because the trials have not been properly designed. Too often, people with normal to high 25OHD concentrations are enrolled and given a small amount of vitamin D. The proper way to conduct such trials was outlined recently in this paper: Heaney RP. Guidelines for optimizing design and analysis of clinical studies of nutrient effects. Nutr Rev. 2014 Jan;72(1):48-54.

Disclosure I receive funding from Bio-Tech Pharmacal (Fayetteville, AR) and Medi-Sun Engineering, LLC (Highland Park, IL).

On 2014 Nov 25, Robert Eibl commented:

Sounds like a very good step forward in understanding these cells.

On 2015 Jan 12, Larry Parnell commented:

This paper was my selection for Paper of the Week for 10-14 Nov 2014.

On 2015 Mar 22, Sean Ekins commented:

And here is a post on the letter from editor http://www.collabchem.com/2015/03/22/what-warrants-an-erratum-and-why-the-old-publishing-model-must-change/

On 2015 Mar 22, Sean Ekins commented:

Here is the letter CAS sent to J Med Chem Editor (which was forwarded to me) regarding this paper http://figshare.com/articles/jm_2014_011308_CAS_report_of_error_2_pdf/1348274 alerting us to 20 document records for one patent which were not visible to us when the analysis was done in 2014.

On 2015 Feb 11, Christopher Southan commented:

None

On 2014 Dec 17, Sean Ekins commented:

Here are some slides on this work http://www.slideshare.net/ekinssean/medicinal-chemistry-due-diligence-computational-predictions-of-an-experts-evaluation-of-the-nih-chemical-probes

On 2014 Dec 17, Sean Ekins commented:

Here is a link to a post discussing this paper and also link to in the pipeline mention https://www.collaborativedrug.com/buzz/2014/12/16/beware-researchers-challenges-navigating-commercial-and-public-databases/

On 2014 Dec 15, Sean Ekins commented:

Also added in my summary of 2014 papers http://www.collabchem.com/2014/12/15/a-year-in-publications-2014/

On 2014 Dec 04, Sean Ekins commented:

Here is a link to a discussion of the peer review of this paper and it includes all reviewer comments http://www.collabchem.com/2014/12/04/chemical-probes-and-parallel-database-worlds-who-wants-to-know-more-publishing-fun/

On 2014 Dec 02, Nadia Litterman commented:

Here is a link to the related paper on using maching learning to model an expert's evaluations and due diligence: http://www.ncbi.nlm.nih.gov/pubmed/25244007

On 2014 Nov 29, Christopher Southan commented:

Comparative statistics updated and live links added at http://cdsouthan.blogspot.se/2014/11/pmid-25415348-back-story-on-bioactivity.html

On 2014 Nov 28, Paolo Tieri commented:

This article is one of the "Multi-omic Data Integration" Research Topic in Frontiers, which focuses on data integration approaches and methods of any type and extent, their application in understanding the pathogenesis of specific diseases or in identifying candidate biomarkers, in order to exploit the full benefit of multi-omic datasets and their intrinsic information content. For more information about multi-omic data integration check http://journal.frontiersin.org/ResearchTopic/2280