On 2017 Oct 27, Stuart RAY commented:

Is it absolutely clear that one can accurately infer "haplotype frequency counts" after PCR and 454 sequencing? As the Lorenzo-Redondo manuscript notes, the template concentrations are low.

Liu SL, 1996 noted that template resampling is likely under such conditions; it's not clear to me that "clean up" procedures resolve this. "Controversial" seems appropriate - so does "intriguing", and I look forward to the next chapter in this vital story.

On 2017 Oct 06, Steven M Wolinsky commented:

In our view, the contrarian findings by Kearney et al. on the sequence data from our paper are entirely due to severely biased subsampling of the data (ignoring haplotype frequency counts), and inappropriate rooting of trees that can give rise to profoundly misleading inferences of lineage relationship and viral evolution. There is evidence of forward evolution for a notable subset of patients studied by many of the same authors as in Kearney et al. using single genome sequencing when rooting the trees on an appropriate outgroup. The study of Van Zyl et al., while clearly valuable and interesting, does not directly contradict our findings because: it only examined sequences from a single compartment (bloodstream); did not plumb sequences deep enough to ensure reliable detection of low-frequency viral variants; and used an entirely different study population.

On 2017 Sep 20, Stuart RAY commented:

The findings described here are controversial. Others who analyzed the same sequence data did not reach the same conclusions - and some of the latter report's authors are also involved in another recent study suggesting evolutionary stasis in children receiving suppressive antiretroviral therapy Van Zyl GU, 2017. It will be important to address these discrepancies, because they address issues central to treatment efficacy and HIV cure.

On 2017 Sep 20, Stuart RAY commented:

The findings described here are controversial. Others who analyzed the same sequence data did not reach the same conclusions - and some of the latter report's authors are also involved in another recent study suggesting evolutionary stasis in children receiving suppressive antiretroviral therapy Van Zyl GU, 2017. It will be important to address these discrepancies, because they address issues central to treatment efficacy and HIV cure.

On 2017 Oct 06, Steven M Wolinsky commented:

In our view, the contrarian findings by Kearney et al. on the sequence data from our paper are entirely due to severely biased subsampling of the data (ignoring haplotype frequency counts), and inappropriate rooting of trees that can give rise to profoundly misleading inferences of lineage relationship and viral evolution. There is evidence of forward evolution for a notable subset of patients studied by many of the same authors as in Kearney et al. using single genome sequencing when rooting the trees on an appropriate outgroup. The study of Van Zyl et al., while clearly valuable and interesting, does not directly contradict our findings because: it only examined sequences from a single compartment (bloodstream); did not plumb sequences deep enough to ensure reliable detection of low-frequency viral variants; and used an entirely different study population.