On 2021-09-08 13:37:06, user Santiago Justo Arevalo wrote:

Article published in Scientific reports of Nature: https://doi.org/10.1038/s41...

On 2021-09-08 10:00:05, user Varun Kapoor wrote:

Hi, Nice article. I really enjoyed reading it and very much liked the idea about using slice by slice segmentation and then using tracking to create a 3D object out of it.

I just wanted to point out that in your references you cite from [7-11] all the articles that made deep learning based tracking and or track analysis softwares in Fiji/Napari. We did work in the same direction and published it in July in the proceedings of the scientific python 2021. Could you please also cite our work along with the other works you cite? The link to our paper: http://conference.scipy.org...

DOI: 10.25080/majora-1b6fd038-014<br /> The bibtex and pdf of the paper are available on the above link.

Many Thanks,<br /> Varun Kapoor

On 2021-09-08 08:19:11, user Rony wrote:

Published version: https://doi.org/10.3390/ijm...

On 2021-09-07 17:14:38, user Rohit Ruhal wrote:

Need screening of chemical library

On 2021-09-07 15:15:05, user Manickam Lab wrote:

This paper is now published in the Journal of Controlled Release: https://www.sciencedirect.c...

On 2021-09-07 15:14:27, user Manickam Lab wrote:

https://www.sciencedirect.c...

On 2021-09-07 14:57:52, user Pankaj Kumar wrote:

The preprint has been published and a link will be forthcoming. In the accepted version the primer set has been updated due to an error in the preprint version.

On 2021-09-07 00:37:51, user Monsif Shawky wrote:

Thanks to Dr. Osman she didn’t accept any unethical behavior accompanied with the submission of this manuscript and she refused to accept being a first author on a manuscript was developed and created by Mr. Shawky.

SNPs derived from the breast cancer GWAS was mapped to human genome 38 while rao et al Hi-C were mapped to human genome 37.<br /> I believe Mr. Shawky performed leftover for the SNPs and mapped them to human genome 37

This is the first manuscript to show the huge impact of the 3D structure of the genome on SNPs functions

On 2021-09-06 16:53:52, user Cedric Berney wrote:

Great to see these new data on sanchytrids!

I was wondering if you did consider the possibility that sanchytrids are actually branching inside Blastocladiomycota.<br /> There is no transcriptomic/phylogenomic data available yet for any member of the shorter-branched Physodermatales.<br /> Therefore there is a possibility that they would turn out to be basal to Blastocladiales + sanchytrids, making it premature to create a new phylum for sanchytrids.

Also there is very strong support for Blastocladiales + sanchytrids in the tree (the internal branch is actually longer than that at the base of Chytridiomycota, Zoopagomycota and Mucroromycota).<br /> So to be phylogenetically consistent across Fungi, one could argue that either the latter three should become multiple phyla, or Blastocladiales + sanchytrids should be in a single phylum, irrespective of the latter's unique traits.

In any case, whether sanchytrids deserve to be a separate phylum or not, the very strong support for their relationship to Blastocladiomycota means that this B+S clade is evolutionarily relevant and would arguably deserve a name, like Dikarya for Ascomycota + Basidiomycota + Entorrhizomycota.<br /> Any suggestion for that name?

Best wishes

On 2021-09-06 16:07:11, user son vi wrote:

Great idea. However, in Supp Figure 18, authors showed all the new versions are extremely prone to giving false positives which challenge the use of such modified enzymes for diagnostic test ? (Supp Fig 18: at 72 degree C Non Template Control can have Ct thresold as soon as 10-14 minutes; and reach full signal after appx 20 minutes; at 74 degree C the spurious activity was reduced but still the system will be less robust if one relies on controlling a 2 degree C difference)

Can you modify Bst to make it give less suprious non-specific amplification?

On 2021-09-05 15:02:51, user Leo G. wrote:

Does the position S:142 influence any enhancing or neutralizing antibodies? The mutation S:G142D appeared from nowhere, peaked to 60% of all samples, then dropped off.

Also, the virus seems to evolve in an entirely different direction.

For example, S:K417N has almost disappeared

On 2021-08-24 18:10:44, user user wrote:

That data presented in Figure 7D is intriguing. If those sera are still available, it would be useful to know the titers of antibodies against particular spike protein epitopes in each sample.

On 2021-09-05 15:00:48, user UAB BPJC wrote:

Review of Barrasso et al., “Impact of a human gut microbe on Vibrio cholerae host colonization through biofilm enhancement” by the University of Alabama at Birmingham Bacterial Pathogenesis and Physiology Journal Club

Summary

This lab previously showed that Paracoccus aminovorans could be found in higher abundance in Vibrio cholerae-infected individuals compared to non-infected individuals. This study demonstrates that V. cholerae colonization is increased by the presence of P. aminovorans in an infant murine intestinal model of infection. Through crystal violet staining and murine intestinal colonization with a ∆vpsL mutant of V. cholerae, it is shown that Vibrio exopolysaccharide (VPS) is necessary for P. aminovorans-dependent enhancement of V. cholerae biofilm formation and intestinal colonization. Microscopy also reveals VPS enrichment in areas of the pellicle with higher P. aminovorans abundance. Lastly, with mutants in accessary matrix proteins RbmA, RbmC, and Bap1, the researchers show that the ability of V. cholerae to form a structurally intact biofilm is necessary for P. aminovorans-dependent enhancement of V. cholerae colonization.

Overall, this is a very interesting paper with insightful experiments that sparked a great discussion in our journal club group. This paper gives strong evidence that P. aminovorans promotes V. cholerae colonization in a VPS-dependent manner. With that said, we have some comments that may be beneficial for the authors to address.

General comments<br /> * May be beneficial to keep y-axis scales for CFUs the same among all figures for more consistency<br /> * One-way ANOVA may be a more appropriate test for figures in which multiple comparisons are being made; we advise you to consider consulting a statistician on the appropriate statistical tests to use<br /> * The Mann Whitney U test is not a t-test, but is referred to as such in some of the figure legends. <br /> * It is mentioned multiple times that crystal violet absorbance was measured at 570nm, although measurements were made at 550nm for all crystal violet figures in the paper. <br /> * Discuss possible limitations with Wheat Germ Agglutinin; could WGA also stain GlcNAc produced by P. aminovorans?<br /> * Would be beneficial to describe in more detail the purposes of matrix proteins RbmA, RbmC, and Bap1, and why these were chosen to be studied.<br /> * One statistical test is mentioned at the end of every figure legend; is the same statistical test being formed on all panels in a given figure? If not, this needs to be clarified.<br /> * More details about how CFUs are quantified from pellicles in the results section would be good to add; the steps taken are only briefly mentioned and are a bit difficult to understand.

Figure-specific comments<br /> * Story may flow better if in vivo data from Figure 2 is added to the end of the paper along with Figure 7 in vivo data; introduce enhanced Vc colonization phenotype with in vitro data first<br /> * Would be beneficial to show single and dual species P. aminovorans CFUs in Figure 2B and 2C and also single species P. aminovorans CFUs in Figure 3B to see whether P. aminovorans colonization also increases in the presence of V. cholerae <br /> * Address the purpose of the grid in Figure 4A<br /> * Figure 5A: Title claims that P. aminovorans increases V. cholerae biofilm formation according to data in Figure 5, but can only make this claim if CFUs are quantified in the in vitro model.<br /> * Figure 6 should have a Vc single species control to compare dual species pellicle to.

On 2021-09-05 14:38:51, user Rodrigo Lorenzi wrote:

I just took a look at the article. My question is: what happens when someone vaccinated is infected by a variant. Do they produce new antibodies against this variant or the only antibodies at work are those induced by the vaccine?

On 2021-09-04 20:28:34, user peter bayley wrote:

for me the most puzzling aspect of this paper is that nowhere is it reported (or even speculated upon) whether plasma treated rats lived longer than either control rats or the average age of the rat strain used in the experiment. That would seem to be the acid test of these types of experiment. Clinically it doesn’t matter if a biological marker shows a “healthier” profile if the health or longevity of the organism, whether rat or human, doesn’t change.

On 2021-09-04 13:43:02, user Arkom Chaiwongkot wrote:

This paper is now accepted and published in PLOS ONE, can be reached via link https://journals.plos.org/p...

On 2021-09-03 19:05:45, user Christopher Muir wrote:

After review, we decided to collect more data and reanalyze. This took awhile and resulted in a quite different paper. Hence, we have decided not to revise this manuscript further and no revisions will be made. Please see this preprint that addresses similar ideas with a much broader data set:

https://www.biorxiv.org/con...

On 2021-09-02 20:36:00, user Célio Dias wrote:

The link supplied to the code just does not work

On 2021-09-02 07:24:36, user Max Gattie wrote:

This article has been split into two.. The first half is available (open access) at Frontiers in Integrative Neuroscience. https://www.frontiersin.org...

On 2021-09-01 19:27:42, user Chen Sun wrote:

This paper developed a universal cryo-EM fiducial for the study of protein-nanobody complexes and validated with two membrane proteins. The only concern is, the process of producing nanobody for membrane protein is long and expensive. A minor point is that the mask used for SFig.3f FSC is obviously too tight.

On 2021-09-01 13:41:17, user Isabel Cardoso wrote:

Hi, Our preprint has been published and a link will be forthcoming.

On 2021-09-01 08:57:30, user Dong-Ha Oh wrote:

Supplementary Figure S1-S11 can be found within the main PDF. They are presented at the end of the PDF, together with the main Figures in the order of appearance in the text. In the "Full Text" tab, they are at the end of the document.

On 2021-09-01 00:46:56, user Tom J wrote:

Please check Fig 1d relative to the text on page 4...seems contradictory. If the Alpha spike/Delta backbone replicated less efficiently than Alpha, the ratios should be >1. Is the y-axis in the figure correct?

On 2021-08-24 22:09:46, user Kelly Ten Hagen wrote:

Our recent preprint provides mechanistic insight as to why P681 is important. P681 enhances O-glycosylation in this region, which blocks furin cleavage. Mutation of P681 abrogates O-glycosylation and leads to increased furin cleavage.<br /> https://www.biorxiv.org/con...

On 2021-08-31 14:27:55, user Greg Keele wrote:

The final peer-reviewed paper is now available from Cell Genomics.

https://www.cell.com/cell-g...

On 2021-08-31 09:48:34, user Mimmo wrote:

Figure 3 is too small to see anything...

On 2021-08-30 17:49:42, user David Quain wrote:

On the one hand, I was pleased to see this preprint. On the other hand I was surprised and disappointed to see no recognition of the work published in peer reviewed brewing science journals on draught beer microbiology/quality.

The Journal of the Institute of Brewing was first published in 1895 and receives just one reference (bizarrely to the Institute of Brewing & Distilling). To my knowledge, Seton in 1912 discussed draught beer quality in the trade with publications in the 50s, 70s and 80s.

In recent years, we have published a review on the microbiological quality of draught beer (2015), the hygiene of tap nozzles (16), assessment of quality (18), survey of trade quality (19) and assessment of biofilm formation (21). None of these publications merit consideration/recognition in the preprint by Bose et al. I'm hopeful that the peer review process will identify these gaps.

On 2021-08-29 13:51:27, user Maryam Fouladvand wrote:

This is <br /> really a nice work. During skeletal muscle regeneration, cells' shape is really different from normal one. Using MyoView, help us to save time and get more solid results. well done

On 2021-08-26 23:28:45, user Laura Sanchez wrote:

Dear Willems et al, this preprint was discussed in a lab meeting and we would like to offer the following for review. Thank you for posting this very interesting manuscript. Best, The Sanchez Lab:

The manuscript by Willems et al. presents an overview of a new software, AlphaTims, which allows for fast indexing and retrieval of LC-TIMS-MS/MS data. The authors present a clear explanation of how software indexing and data matrix construction works, along with an example of how their software is used to simplify the accession of data from complex samples. This open-source program has the potential to make data more accessible without proprietary software, and has many possible applications in conjunction with further data processing software. Overall, AlphaTims looks to be an impressive piece of software, and figure 1 especially did a great job of visualizing and explaining the difficult concept of data indexing in 4 dimensions. This was a great, polished read, and very informative in elucidating the experimental process for those who are less familiar with the LC-TIMS-MS/MS workflow. It was also appreciated that the data was publicly available on PRIDE! Below please find a list of critiques that may improve the accessibility of some of the concepts and better illustrate the software’s utility.

● Figure 1 is very well done and all the terms are well-explained. It may add further clarity if the authors were to add a part (d) to help visualize what the data looks like once indexed, perhaps with number values. Additionally, the color palette is difficult to interpret when printed in black and white; the authors may wish to consult ColorBrewer.

● Some of the data on the speed of the software is difficult to conceptualize without benchmark data for comparison. There’s a brief comparison between the program and Bruker’s DataAnalysis on two different systems (one local and one Github VM) with different specifications, which doesn’t provide a lot of context as these do not seem to be directly comparable. While the speed of the software compared to others is not necessarily a major focal point, as AlphaTIMS carries out different processes to DataAnalysis, it’s still a little difficult to conceptualize the relevance of the times in figure 2, as these events are somewhat decontextualized. With that being said, it was apparent that both systems used solid state drives. Therefore, even if runtimes cannot be directly compared, it may be helpful to note whether the code is CPU-bound, storage speed bound, etc. and suggest what hardware upgrades may help increase performance.

● The documentation in general is very good! The accessibility and the included explanations for all of AlphaTims’ reading, writing, slicing, and visualization workflows are impressive. However, more examples for other included processing functions (i.e. centroiding) would be helpful.

● Some more broad data visualizations in figures 3 and 4 would be helpful. Seeing an overall LC-MS and/or TIMS-MS spectrum in figure 3 could help contextualize the complexity of the data, beyond the quality control purposes expressed in the figure. This would also help to visualize the described “polygon filter” - again, not an expressed priority of the paper, but helpful for readers to connect the software to data. Additionally, in figure 4, showing a comparison of the selected peptides between the 6 sample conditions could help to visualize the utility of the software for determining sample optimization. There’s also a slight disconnect in the captions for figure 4 - “cell 24” could refer to In [24] or Out[24]. This is worth clarifying with traditional a, b, c, d, e, f labels.

● The forward looking statement in the conclusion may provide a great opportunity to expand on future possible applications. For instance, do the authors see this being developed with additional data processing/analysis functionalities, or integrated into new/existing data analysis programs? Is there possible functionality to visualize and compare multiple datasets at once?

On 2021-08-26 12:39:17, user Kirk Overmyer wrote:

This paper has now been published in the journal Forest Research and can be accessed here:

http://www.maxapress.com/ar...

Best Regards,<br /> Kirk Overmyer

On 2021-08-26 08:36:48, user zdz wii wrote:

Where is the supplementary material?

On 2021-08-26 08:17:05, user Brian wrote:

hello can you add data/code here?

On 2021-08-25 17:35:35, user S wrote:

The Pfizer and Moderna vaccines CAN neutralize the Lambda variant! Johnson & Johnson might be less effective but can also neutralize it too!

https://www.biorxiv.org/con... if they got J&J)

On 2021-08-25 15:49:58, user Julie Hanson Ostrander wrote:

An Open Access, read-only version of the peer-reviewed article published in Oncogene can be accessed through the following link: https://rdcu.be/cv1jC . The Open Access, published article can also be accessed through the J. Willard Marriott Digital Library at the University of Utah: https://collections.lib.uta...

Additional disclosures found in the published manuscript.

This work was supported by NIH grants R01 CA236948 (JHO, CAL), R01 CA229697 (CAL), F32 CA210340 (THT), T32 HL007741 (THT), U54 CA224076 (BEW), R01 CA248158-01 (CODS), and R01 AG069727-01 (CODS). ACS Institutional Research Grant #124166-IRG-58-001-52-IRG5 (JHO), University of Minnesota Masonic Cancer Center (CAL, JHO), the Tickle Family Land Grant Endowed Chair in Breast Cancer Research (CAL), National Center for Advancing Translational Sciences of the NIH Award UL1TR000114 (JHO), and Department of Defense W81XWH14-1-0417 (BEW). We thank Bruce Lindgren for biostatistics support, and the Masonic Cancer Center Biostatistics and Bioinformatics, Analytical Biochemistry, University Imaging Core (UIC), and Flow Cytometry cores. We also thank Zohar Sachs and Michael Franklin for critical reading of this manuscript.

We disclose that CAL is a Scientific Advisory Board Member for Context Therapeutics, Inc. BEW, EC-S, KPG, and C-HY may receive financial compensation from intellectual property and tangible property licenses managed by the University of Utah. The remaining authors have no COI to disclose.

On 2021-08-24 19:42:40, user Julie Hanson Ostrander wrote:

Use this link to access a view-only version of the article for free and use Enhanced PDF features such as annotation tools, one-click supplements, citation file exports and article metrics. https://rdcu.be/cv1jC

On 2021-08-25 13:27:36, user Marc Scheetz wrote:

This has now been published in AAC. https://journals.asm.org/do...

On 2021-08-25 12:52:36, user Sergio Perez Acebron wrote:

Revised manuscript published in PNAS: https://www.pnas.org/conten...

On 2021-08-25 07:27:33, user nemo peeters wrote:

Dear Readers, A colleague has spotted a copy/paste error resulting in a yeast spot duplication in Fig1A. The new figure will be uploaded soon, together with a new supplementary material showing the raw Y2H matrices.

On 2021-08-24 23:52:42, user george mcnamara wrote:

I hope the authors read and cite Duong et al 2013 CAR library approach, before their next version.<br /> https://journals.plos.org/p...

On 2021-08-24 22:13:47, user Kelly Ten Hagen wrote:

Recent work from our lab provides mechanistic insight as to why P681 is important. P681 enhances O-glycosylation in this region, which blocks furin cleavage. Mutation of P681 abrogates O-glycosylation and leads to increased furin cleavage.<br /> https://www.biorxiv.org/con...

On 2021-08-24 22:05:29, user Kelly Ten Hagen wrote:

Our recent preprint under review provides mechanistic insight as to why P681 is important. P681 enhances O-glycosylation in this region, which blocks furin cleavage. Mutation of P681 abrogates O-glycosylation and leads to increased furin cleavage.<br /> https://www.biorxiv.org/con...

On 2021-08-24 17:44:27, user Bogdan Pasaniuc wrote:

Many thanks for the very insightful questions and the super neat related literature (as usual the breeding genetics world has made super insightful advances prior to the human genetics community!). Your comments will allow us to clarify the points below and significantly improve the quality of our manuscript!

Re Q1:<br /> Yes indeed we use standardized genotypes for the purpose of having a simple toy figure. The main point we try to convey is that multiple causal effect size configurations can lead to the same observed marginal effects in GWAS. Thus, given GWAS data only, our approach proposes to sample across these causal configurations to estimate heritabilities. As with all toy figures, there is a balance between oversimplification and leading to misinterpretations; we will clarify this better in the legend/text.

Re Q2: <br /> First, indeed we make the assumption that the true causal effects \beta are independent across SNPs; this is a standard assumption that is made across most heritability work in human genetics and likely a good approximation in real data. That being said, we can drop this assumption then an extra covariance term exists (see bottom pp17) that could potentially be estimated/investigated; here we focus only on the \beta’R\beta term (our estimand of interest).

Second, as you clarify and we are in full agreement, the posterior samples have a covariance structure that is different from identity; i.e. \beta_i and \beta_j post samples are correlated. In the most simple case of the toy example of Fig 1 with two SNPs in perfect LD and with a sparsity model that only allows 1 causal, only one of the \beta’s will have non-zero effect in any configuration; therefore the two betas in the posterior are negatively correlated (r=-1). Or in the case of full infinitesimal model with independence of beta as prior, one can also straightforwardly derive the variance of the posterior as 1/n (1/(1-h2)R + M/(Nh2) I )^-1 (with apologies for self reference see https://www.biorxiv.org/con... or multiple other previous works with similar derivations). In the more general case when there is also a sparsity prior, an analytical solution for the posterior is hard to derive; this motivated us to sample from the posterior of \beta in this work.

Third, as defined in pp 17-18, our estimand of interest is \beta’R\beta where \beta are the unknown causal effects (also denoted as h2gene). We rely on an approximation of the posterior of \beta from SuSiE to sample from posterior effects and then approximate a sampling from posterior of \beta’R\beta (pp19); our proposed estimator has the simple form of avg (\beta’R\beta), where the average is taken across samples from posterior (samples that will have correlations across SNPs, as noted above). We fully acknowledge that other estimators for \beta’R\beta can be proposed (analytical and/or sampling based) that could be potentially more efficient and/or unbiased. In this work we chose to focus on this simple estimator that works reasonably well in simulations and real data.

We will revise to clarify all these points!

On 2021-08-20 09:56:24, user Gregor Gorjanc wrote:

Thanks for this nice work!

I would like to point out that your Figure 1 can be misleading. Additive genetic variance at a single bi-allelic locus under Hardy-Weinberg equilibrium is 2pq\beta^2, not just \beta^2, where p and q are allele frequencies and \beta is allele substitution effect. In general, this is 2pq(1+F)\beta^2, where F measures deviation from Hardy-Weinberg equilibrium. So, genetic variance and heritability at a locus depend on variance of genotypes and allele effects (classic reference for this is Falconer and MacKay, 1996). Hmm, are you skipping variance of genotypes, Var(x), because you standardise your genotypes (which changes what \beta mean)?

When you expand to a region, linkage-disequilibrium between loci within the region also kicks in. Also the region will be correlated to other regions, so variance partitioning is not really that meaningful looking just one region or locus at a time. You account for linkage-disequilibrium with your R matrix. But, when working with regions, correlation between estimates of \beta_i and \beta_j also kick in, even though we assume apriori that \beta_i and \beta_j are independent. Your derivation on page 17 works on the prior side, but does this also hold when you work with posteriors of \beta - they are quite correlated and possibly also correlated with genotypes (x) too?

We have been doing something similar, but focusing on the whole genome and partitioning additive genetic variance by chromosome (=region of a genome) https://www.biorxiv.org/con.... This work of ours relied a lot on the past work of:

Sorensen, D., R. Fernando, and D. Gianola, 2001 Inferring the trajectory of genetic variance in the course of artificial selection. Genetics Research 77: 83–94.

Gianola, D., G. de los Campos, W. G. Hill, E. Manfredi, and R. Fernando, 2009 Additive genetic variability and the Bayesian alphabet. Genetics 183: 347–363.

Lehermeier, C., G. de Los Campos, V. Wimmer, and C.-C. Schön, 2017 Genomic variance estimates: With or without disequilibrium covariances? Journal of Animal Breeding and Genetics 134: 232–241.

Allier, A., S. Teyssèdre, C. Lehermeier, B. Claustres, S. Maltese, et al., 2019 Assessment of breeding programs sustainability: application of phenotypic and genomic indicators to a north european grain maize program. Theoretical and Applied Genetics 132: 1321–1334.

Schreck, N., H.-P. Piepho, and M. Schlather, 2019 Best prediction of the additive genomic variance in random-effects models. Genetics 213: 379–394.

On 2021-08-24 15:53:56, user Pedro Mendes wrote:

Nice work. Table 1 should indicate that COPASI (of which I am one of the authors) is <br /> capable of both fixed-interval output and actual time step output (which<br /> is obtained by selecting the option "automatic" in our time course <br /> settings).

On 2021-08-24 13:34:31, user Georg Pabst wrote:

The paper has been published (open access) by now. Go to https://doi.org/10.1016/j.b...

On 2021-08-24 10:28:42, user yuri wrote:

Very nice work and well done. Was wondering why using adaptive sampling and not 16S with the rapid kit from ONT. Was it for comparison with Illumina? In case, I'd just give it a try, it'd cost one kit or a couple of primers + LSK-109 and a single flowcell. Just a curiosity though. It's also help doing an unbiased analysis of bacteria, just in case you missed something new....

On 2021-08-24 08:32:04, user Steven Groot wrote:

This is a very interesting manuscript, showing the potential of low oxygen storage for prolonging seed life span, while pointing to an important critical control point. Moreover, the manuscript provide important data on the biology behind.

On 2021-08-23 15:26:46, user Crap wrote:

Elimination of flow-through - does the Receptor not breathe?<br /> 50% efficacy study was done only for coughing, and not breathing, in an environment for only filtration through the mask, not taking into account side air flow-through. The study, although controlled, has questionable results at best. <br /> Further, model assumed an open room in calculations - eliminating any barriers, and heat sources for impact on mass transfer volumetric flow.

On 2021-08-23 00:57:39, user Martin Frith wrote:

Thanks for advancing our understanding of this interesting topic!

1. The definition of W4 has a typo: line 4 duplicates line 2.

2. In Fig. 4a it would be interesting to see ry words, which have the same d=1/4, instead of (a,b,n)-words.

3. Context-dependency decreases sensitivity, but it also increases specificity, does that compensate?

On 2021-08-22 21:06:38, user Fabiane Matos dos Santos wrote:

The research article entitled "Aspirin-triggered resolvin D1 reduces parasitic cardiac load by decreasing inflammation through N-formyl peptide receptor 2 in a chronic murine model of Chagas disease" by Ileana Carrillo et al., investigates the role of Aspirin-triggered resolvin D1 (AT-RvD1) as a pro-resolving mediator of inflammation that acts through N-formyl peptide receptor 2 (FPR2) using a FPR2 knockout murine model of chronic Trypanosoma cruzi infection. The data provided in the manuscript is important and interesting in connecting a link between AT-RvD1 as an attractive therapeutic activity due to its regulatory effect on the inflammatory response at the pathophysiology to cardiac pathology during T. cruzi infection. The manuscript is an original research article and it suffers with the following minor concerns:

Materials and methods section<br /> 1. Explain the rationale for the use of Dm28c strain of T. cruzi. What is the rationale to select this strain and provide more data concerning its virulence and susceptibility level to benznidazol?

Results and discussion sections<br /> 1. Include the data of eletrocardiographic variables in the manuscript, even in the absence of diferences among the groups of mice. Discussion about eletro<br /> cardiography abnormalites in Chagas cardiomyopathy needs to explore more parameters than only QT interval, for example: PR interval (iPR) and atrial block, axis of QRS, evidences of atrial fibrillation and right bundle branch block.

On 2021-08-21 03:53:31, user Aneth David wrote:

I found this paper really interesting and comprehensive. I liked the fact that they studied fungal communities too, I hadn't come across any study that has done this for reasons I don't understand.

I thoroughly enjoyed reading this paper.

On 2021-08-19 19:16:55, user Patricia G Cipriani wrote:

Please be aware that our article describing the characterization of interactions, phenotypes and localization of MIP-1 (C38D4.4) and MIP-2 (F58G11.3) has been published on July 5th, 2021. https://elifesciences.org/a.... These names were officially approved by the WormBase curator on July 12th, 2019.

On 2021-08-19 14:35:18, user Meng Wang wrote:

We have recently reported that the Tn5-based epigenomic profiling methods, especially Stacc-seq and CoBATCH, are prone to open chromatin bias (https://www.biorxiv.org/content/10.1101/2021.07.09.451758v1). Rather than directly address this bias issue, the authors of Stacc-seq argued in this preprint that FC-I normalization (normalizing by input/IgG control) was better than FC-C (normalizing by background) for Stacc-seq etc. data analysis. Based on this, they claimed that our results had “a major analysis issue”. However, the truth is that we had already used both FC-I and FC-C normalization methods and both showed clear open chromatin bias for Stacc-seq and CoBATCH. The fact that our analyses demonstrating that CUT&Tag (5% FPR) showed much lower FPR than Stacc-seq (30% FPR) or CoBATCH (50% FPR) indicated that the high FPRs were not due to “artificially enhanced the relative enrichment of potential open chromatin bias”, but an intrinsic problem of Stacc-seq and CoBATCH. In our opinion, the preprint has several problems, which are detailed below.

1. The preprint ignored the fact that we had already used both FC-I and FC-C normalization methods. The authors assumed that we only used FC-C for Stacc-seq etc. (Fig. 1A in Liu et al.). However, in fact we used both FC-C and FC-I in our analyses. In Fig. 1c, d and Fig. S2 of our manuscript (Wang et al.), methods labeled with “with IgG” were results from FC-I normalization, and methods without such label were results from FC-C normalization. Importantly, results from both normalizing methods showed clear open chromatin bias for Stacc-seq and CoBATCH (Fig. 1c,d and Fig. S2 in Wang et al.).

2. The results of global H3K27me3 enrichment at the Polycomb targets in this preprint (Fig. 1C) was contradictory to their claim that using FC-C would cause “complete loss or dramatic reduction of enrichment at true targets for datasets generated by Tn5-based methods”. Fig. 1C of this preprint showed a clear H3K27me3 enrichment around the TSS of Polycomb targets compared to adjacent regions when using FC-C. The difference between results from FC-I and FC-C is caused by the y-scale. The fold change is a relative measurement, so the y-scale of different normalization methods is not directly comparable. If they set the y-scale of FC-C to 0~2, the enrichment pattern would be highly similar to that using FC-I.

3. The genome browser snapshots of several loci in a large scale (low resolution) could not demonstrate that the results from FC-I and FC-C normalization are globally different. This preprint provided several example loci (Fig. 1B and Fig. 2 in Liu et al.) to show that using FC-C would cause “complete loss or dramatic reduction of enrichment at true targets for datasets generated by Tn5-based methods”. However, showing browser view of very large regions are misleading as the resolution is too low. For genome browser display, the look of the signal track patterns depends on y-scale, x-scale and windowing and smoothing function. When viewing a very large region, the signals are sampled and aggregated by genome browser and are not the raw signals. Thus, the patterns may not reflect the real situation. Indeed, when zoomed-in to check these regions, we found the peak patterns from FC-I and FC-C normalization are highly similar. In addition, examples from several loci could not reflect the global pattern. The global enrichment shown in Fig. 1C of this preprint did not support their conclusion, as discussed in point 2.

In summary, our original analysis has already included the normalization method suggested by the authors of this preprint. Results from both normalization methods supported that Stacc-seq and CoBATCH had high open chromatin bias. In fact, the results from this preprint also support our conclusions. In Fig. 2 of this preprint, regardless whether FC-I, FC-C or RPKM were used, the discrete peaks from Stacc-seq etc. were more similar to ATAC-seq peaks, but were totally different from ChIP-seq peaks.

Meng Wang and Yi Zhang<br /> Howard Hughes Medical Institute, Boston Children’s Hospital, Boston, Massachusetts 02115, USA

On 2021-08-19 13:16:56, user Michael Galko wrote:

Losick: “A previous study reported cell enlargement with loss of cell-cell junctions in the adult Drosophila ventral abdominal epithelium with age (Scherfer et al., 2013), but did not characterize the extent of multinucleation nor whether cells were becoming polyploid.”

Scherfer 2013: Fig 1D (3 days old) and especially Figure 1E (14 days old) show clear multinuclearity of the aging epidermal cells that progresses with age. Fig 2B shows by transmission EM multinucleate cells that have no intervening membranes or junctions.

It is cool that the fusion can be blocked and that Brainbow analysis<br /> confirms the continuous cytoplasm. That said, proper credit for the first<br /> observation of this multinuclearity should be given. The paper here is written as if the Scherfer study did not observe/report multinuclearity. This is not correct.

On 2021-08-18 05:58:07, user Milan Vrtílek wrote:

This work has been just published under title "Macroevolutionary foundations of a recently evolved innate immune defense" in Evolution https://onlinelibrary.wiley...

On 2021-08-17 20:22:55, user Matt wrote:

Hi authors; regarding the findings of Siberian ancestry patterns in Europe (which roughly seem to conform to a distance based pattern from the northeast to southwest), do these replicate in simpler f statistic measures? E.g. the same clinal patterns are present and they replicate without an explanation that can be found in varying proportions of ancient populations within Europe ("WHG" / "EEF" / "Steppe"). If not, isn't this a problem for the finding and if not a problem, why not?

On 2021-08-17 20:14:14, user Sue Bello wrote:

There appears to be something wrong with the pdf. When I attempt to open the downloaded file with Acrobat I get an error message "There was an error reading this document (14)".

On 2021-08-17 16:08:17, user Graham Hitman wrote:

This MS is now published in BMC endocrine disorders <br /> The link is<br /> https://rdcu.be/ct4S2

On 2021-08-17 12:11:04, user passanger lost wrote:

Not found the Supplementary Table files

On 2021-08-17 09:17:54, user passanger lost wrote:

Can not found the SUPPLEMENTARY INFORMATION files

On 2021-08-16 17:36:20, user Frank William Soveg wrote:

Figure 2 and the supplementary materials are missing from the manuscript. Can the authors please make this available? Thank you.

On 2021-08-16 15:57:24, user bric hard wrote:

Very Interesting.I did not expect the results! Take good care Joao...

On 2021-08-16 13:09:27, user Lukas Tanner wrote:

Dear authors, <br /> Thanks for this very interesting study. Similar findings have been previously described by Tabata et al in 2017. They show the same mechanism/principle for thymidine phosphorylase (https://www.cell.com/cell-r... in fueling cancer cells under nutrient deprived conditions.<br /> Thanks and best wishes, <br /> Lukas Tanner

On 2021-08-16 11:00:28, user Gregory Bix wrote:

This preprint has been published in the journal Life Sciences.<br /> https://www.sciencedirect.c...

On 2021-08-16 09:39:34, user Richard Smith Lab wrote:

Supplemental movies:<br /> https://youtu.be/q6TGjKT4lxc<br /> https://youtu.be/l-avOqrB_TI<br /> https://youtu.be/KwrwHor5p_I<br /> https://youtu.be/Bh5QEFjL1Uo<br /> https://youtu.be/31IjUtpP6os<br /> https://youtu.be/jEctM96Algo<br /> https://youtu.be/NyTcEjorXdc

On 2021-08-16 08:53:58, user Qiao-Ping Kevin Wang wrote:

This is a very interesting and elegant study. Lot of GWAS studies have revealed that a few variants like rs1421085 in FTO non-coding region are highly associated with obesity. However, none of these variants has been functionally validated in vivo due to tremendous difficulties and huge costs. In this study, the authors firstly demonstrated that a single rs1421085 variant T>C in the non-coding area was a functional point and the mutation from T to C could boost thermogenesis in global knock-in and BAT specific knock-in mice. Mechanistically, T>C mutation could enhance FTO gene expression, which in some way, boost UCP1 expression. Furthermore, based on the frequency of this mutation in different population around the world, the authors proposed a very intriguing hypothesis that this mutation with increased thermogenesis capacity help human to adapt to coldness, thus in certain way facilitating human to migrate out Africa and spread all the world. This study gives us some fresh view that the SNPs identified from GWAS are functional and that GWAS is an effective way to identify genetic factors for complex diseases.

However, there are some concerns. Since this point localizes in non-coding region, which is usually not conserved across species, how the authors could prove this be true in humans? Except FTO, does this variant affect other known important genes like IRX3 and IRX5, or more thermogenesis related genes? Is possible that the point mutation changes the local chromatin structure if it is so important?

On 2021-08-15 19:28:22, user Anand Ramachandran wrote:

HELLO has been significantly reworked with new methodologies for hybrid and stand-alone variant calling. These methodologies and results are published at:

https://bmcbioinformatics.b...

On 2021-08-15 18:03:00, user Sangeeta Nath wrote:

The paper is now published @ BBA-Mol Basis Dis<br /> https://www.sciencedirect.c...

On 2021-08-15 09:17:23, user David Pellow wrote:

Will the performance of this tool be compared to the standard tools for metagenomic plasmid assembly?

On 2021-08-13 23:33:04, user Xiaoxu Yang wrote:

The manuscript is now published online in Cell https://doi.org/10.1016/j.c....

On 2021-08-13 22:21:40, user Patricia wrote:

Many of your discussion points were addressed in the Lee and Wing preprint, surprised you didn't quote it. https://www.biorxiv.org/con...

On 2021-08-11 21:24:12, user jiarong wrote:

Just want to leave other readers a notice here: I found the accuracy or F1 score of a few gene-based viral identification tools reported in this benchmark study (<0.8) is quite different from those reported in other studies (usually >0.9, e.g. VirSorter2 and VIBRANT). I have got in contact with the authors and took a deep dive into the simulated data, and found there was a significant amount of viral sequence contamination in refseq plasmid databsae and thus in the simulated plasmid contig dataset. There were also some prophage not cleaned by prophage identification tools. The combined viral contaminants in plasmid and host set is comparable to the among of viral contigs in the viral set, which can significantly skew the performance stats. More details can be found in my report here: https://github.com/jiarong/....<br /> I have informed the issue to the authors and they are working on the issue.

On 2021-08-11 06:50:33, user Ticklicker wrote:

Are your planning to share your data? I would like to know the exact date of collection, the county & state, and map coordinates (lat, long) or approximate location (distance & directiion from nearest town) for the 4 "positive' deer samples that were collected in 2019 (1) and January 2020 (3). Also, have you considered that deer samples that tested "positive" for SARS2 may represent instead cross-reaction detection of tick-borne viruses; for example, the "Heartland Virus", which may be spreading eastward since its 2009 discovery at 2 locations in NW Missouri?

On 2021-08-11 02:19:39, user Brian Coyle wrote:

Virus don't easily infect ticks, believe it or not. Coronavirus haven't. It's because virus infect either anthropods or mammals.

On 2021-08-10 16:44:45, user hongmi wrote:

The SARS-CoV-2 sVNT can also detect SARS infection. Actually the cross reaction is stronger 17 years after infection, compared to within a year. Could your positive signal in 2019 and early 2020, and also some low-moderate signals in 2021, was due to this cross reaction?

On 2021-08-03 15:53:33, user Jessica Glasscock wrote:

I was wondering if researchers have considered, or are considering ticks, as the possible link between whitetail deer and humans. I know in the suburban and urban areas of the east coast, deer populations are highly concentrated and lead to increased tick problems and humans being infected with tick borne illnesses. Ticks might be the method in which SARS-CoV-2 was transmitted between humans and deer. Thanks for your time.

On 2021-08-09 13:14:21, user Shuaibing Ren wrote:

Equation 5: P(G) should be P(Gl)

On 2021-08-09 08:24:42, user Jiangming Sun wrote:

All values in a row (black) are 0 in Fig. 1A. Shouldn't remove such row first before any matrix operations?

On 2021-08-09 06:46:31, user Prof. T. K. Wood wrote:

Authors should also compare their directed evolution results with the 20 odd papers that have previously found beneficial mutations on this class of enzymes using directed evolution.

On 2021-08-08 23:50:11, user Chengxin Zhang wrote:

Thank you for sharing this potentially important study. There are a few questions regarding the architecture and the benchmark study.

1. What is the loss function of RGN2 (i.e. loss function for the output of the geometry module). Is it the same as RGN1, i.e. dRMSD, or a new loss function that can distinguish mirrored structures.

2. The performance of RGN1 is highly dependent on the random seed used to initialize the training. Is this still the case for RGN2? In other words, for the same testing protein, if we use two RGN2 models trained on the same training set but different random initialization, what is the GDT_TS between the two resulting structure models?

3. What is the accuracy of predicted theta and phi? This question is similar to a comment on the RGN1 paper (https://www.biorxiv.org/con... which was not fully explored. It is likely that including an auxiliary loss for theta and phi can not only improve local structure correctness but also resolve the chirality issue, which was not addressed by the dRMSD loss.

4. Both RGN1 and RGN2 construct the peptide chain in torsion space, i.e. Ramachandran angles for RGN1; a bond angle phi and a torsion angle theta for RGN2. As pointed out by a recent preprint (https://arxiv.org/abs/2105...., "One well-known problem is that internal coordinates are extremely sensitive to small deviations as the latter easily propagate through the protein, generating large errors in the reconstructed structure." Is there an attempt to address this issue?

5. RGN2 is trained on ASTRAL domains, most of which are "normal" proteins that are neither orphan nor designed proteins. Therefore, the single-sequence structure prediction performance difference between RGN2 and third-party programs (AlphaFold2, RoseTTaFold and trRosetta) should be comparable regardless of whether the testing proteins are orphan or not, provide that only a single sequence rather than MSAs is fed to the three third-party programs. On non-orphan proteins, e.g. CASP14 FM targets that are not used in RGN2 training, can RGN2 still outperform single-sequence-based AlphaFold2, RoseTTaFold and trRosetta?

On 2021-08-07 14:52:11, user Emil Kirkegaard wrote:

Contrary to what the authors say, these data are NOT publicly available. They are in fact hidden and one has to apply via some university institution. https://genomics.ut.ee/en/b...

On 2021-08-07 12:30:57, user Александр Федотов wrote:

There are 0 not predicted proteins, not 29. Please fix it.

On 2021-08-06 17:21:33, user KaAcWh wrote:

Very interesting study. I am curious to hear why the authors did not make any reference to the fact that if Spike is inducing endothelial inflammation via integrin5 and NF-KB, this means that there is a very real risk of such occurrence in people vaccinated with mRNA and vectorized vaccines that lead to the endogenous production of full length (or, even only RBD) of Spike. In other words, 2+2 should equal 4, even in these strange times. Any thoughts?

On 2021-08-05 09:34:08, user William Martin wrote:

This is really interesting. Life on geochemical H2, geochemical formate and geochemical glycine. Just barely possible, but possible, and taking place.

On 2021-08-04 19:35:28, user Jing wrote:

A revised version of this story is now published at -

https://stm.sciencemag.org/...

On 2021-08-04 06:15:38, user Stefan Barakat wrote:

the peer-reviewed version of this paper has now been published in Genetics in Medicine, https://doi.org/10.1038/s41...

On 2021-08-03 19:55:25, user Cornelius Römer wrote:

If you put "higher infectivity" in the title, it requires a comparison. But there's none. Not even in the abstract is the question "what is it more infectious than" answered. Only in the discussion does it become clear that the infectivity is higher than D614G, Gamma and others, BUT NOT Delta.<br /> A better title would be: "SARS-CoV-2 Lambda variant exhibits similar in-vitro infectivity and immune resistance to Delta"

On 2021-08-03 14:27:30, user Hypo wrote:

Why was Illustration g removed?

It was a chart of Percent of CD8 in donor R6 Showing the change in Non Spike to Spike from Post infection to Post Vaccine.

It was a topic of intense discussion among people I know and they found it very interesting.

On 2021-08-03 08:44:06, user Julie Cosmidis wrote:

A peer-reviewed version of this article is now published in Frontiers in Microbiology:<br /> https://www.frontiersin.org...

On 2021-08-02 23:33:12, user Elizabeth Komives wrote:

A critical result from this paper involves the purported observation of bimodal isotopic distributions from HDX-MS data that are interpreted as being indicative of transition to a slowly interconverting form. I would like to point out that quite frequently bimodal distributions are the result of artifacts and the raw data must be carefully examined in order to ascertain their validity. In this case, two other groups have already published HDX-MS data that did not show bimodal distributions strongly suggesting that the results in this paper are due to an artifact of that sample rather than truly indicative of a new state. We would encourage the authors to discuss their results with experts in the field and certainly to show the raw isotopic distributions across the relevant time scale prior to publication.

On 2021-08-02 16:55:23, user SANCHEZ RAYES Ayixon wrote:

Published: Sánchez-Reyes, et al. Hi-C deconvolution of a textile dye–related microbiome reveals novel taxonomic landscapes and links phenotypic potential to individual genomes. Int Microbiol (2021). https://doi.org/10.1007/s10...

On 2021-08-02 16:04:51, user Dr Richards wrote:

The data shown in this publication is INCORRECT and is not endorsed by Charles River. The data shown to represent efficacy of camostat-colloid gold is actually that from the Oseltamivir control. No efficacy was seen with the camostat-colloid gold treatment.

On 2021-08-02 15:55:13, user Maulik Patel wrote:

Proud of Nikita, a graduate student in the lab who conducted the work presented in this manuscript, and overcame several technical challenges over the years. We welcome constructive feedback from the scientific community to help us improve and build on the work in the manuscript. Thank you.

On 2021-08-01 17:38:18, user Abishek wrote:

Our preprint has been published recently. Please find the link to our published work here https://doi.org/10.3390/mic...

On 2021-08-01 14:47:06, user Shrez wrote:

Hi, <br /> Your study looks very interesting.<br /> I was wondering if there is a reason for using GammaP instead of GammaBH from the output of eJTK, when you are selecting cyclical OTUs?<br /> Thanks!

On 2021-07-31 16:18:23, user Ryan Pink wrote:

Excellent piece of work, really interesting, and hats off to you for the detail of the methods...some real integrity here, well done.

On 2021-07-31 14:37:47, user Prof. T. K. Wood wrote:

Would be interesting to compare these, the 7th results showing heterogeneity in persister cell waking, with the 1st report/mechanism based on ribosome dimerization/activity (doi:10.1111/1462-2920.14093, https://doi.org/10.1016/j.i..., doi:10.1111/1462-2920.14828, all not cited).

On 2021-07-31 07:49:48, user pedro chake wrote:

As this paper seems written by a bachelor student (which is a good start!), I would recommend the authors to read the existing literature on the topic, so as to avoid some basic confusions in the terms and objectives. <br /> Most terms are mixed without understanding of their meaning: uncertainty, error, model, map, accuracy, validation. There is a complete mental confusion in this article, and many of the proposed solution arise because of ignorance of the underlying problems and basic misunderstanding of statistical analysis. <br /> Aggregation at the pixel level. Do the authors realize that this is again nonsense? Spatial prediction is made at point, pixels are just visualization. Aggregating (I guess the authors mean averaging) means that the short-scale variation is removed. This is simply wrong and the same as manual modification of the data. This is explained in any spatial statistics book, see the book of Cressie (1991). <br /> For accuracy statistics, see Janssen et al (1995). The paper of Janssen explain what are accuracy statistics, and how to use them. This is basic literature, but still…<br /> Janssen, P. H. M., & Heuberger, P. S. C. (1995). Calibration of process-oriented models. Ecological Modelling, 83(1-2), 55-66.<br /> “spatial insight into the model inaccuracies” is not provided by bootstrapping. This is explained in any book on ecological modelling. Do the authors work in a bubble? And 100 bootstrap is too little to obtain a reliable confidence interval. There are other solutions in the literature, such as quantile regression. But for this the authors should read the literature. <br /> At some point the authors attempt to detect extrapolation, I must say this is poorly made and again there are a lot of confusion of what extrapolation is. I would simply argue that the sampling data should have sufficient coverage to not have to worry about this problem, and if extrapolation occurs the prediction intervals will show it. Also, extrapolation is not simply related min-max of predictor variables. <br /> The authors want to detect extrapolation but then use spatial cross-validation to force extrapolation, that is, validation of the model at unseen places in both geographic and predictor space to correct for spatial correlation. Again, the serious literature never recommended this, can you please read the book of Cressie on spatial statistics and validation of spatial models? <br /> I am sorry to say that also the discussion and recommendation is poor. What about validation of models? Overall in the paper the authors are confused about uncertainty and error, and between the different sources of uncertainty. There are ecological modelling books on this, related to species distribution modelling.

On 2021-07-30 11:03:02, user pedro chake wrote:

It is really worrying for science to see so many of these papers simply denying all what constitute good science. ALL good works are ignored, do the authors really think that they are the first doing mapping and ecological modelling?

On 2021-07-31 05:50:33, user Stephen Fairweather wrote:

The discovery of the essential L-lysine plasma membrane transporter from the human intracellular parasite T.gondii. Large amount of data from various experiment systems and in vivo and very rigorous demonstration of all conclusions. A paper worked on since 2015.

On 2021-07-30 18:01:48, user Andreas Martin Lisewski wrote:

Central to the authors’ conclusions is their Fig. 3A (middle panel) which, however, is misleading: by including the y-axis origin (0) the authors unrealistically imply that they had an unlimited number n of ONT sequence reads. <br /> While apparently they do not report this number for their own experiments, it can be estimated from their Table S1 (n ~ Yield/N50).

It then follows that in Fig. 3A (middle) their data only gives a non-zero upper bound (< 1/n) for the “Proportion of total ONT sequence (%)”, which here becomes about two orders of magnitude smaller than the levels reported by Zhang et al. (to which they directly compare).

In summary, their main message then boils down to the following: without strong over-expression of L1 (as done in Zhang et al.), and without proper controls of other relevant virus/host factors, the fraction of candidate SARS-CoV-2 genomic integration events observed is about two orders of magnitude smaller than the one reported Zhang et al.

This finding certainly does not invalidate (or “contradict") the important and far-reaching result reported by the Jaenisch group.

On 2021-07-30 10:11:56, user David Ron wrote:

This study takes on the interesting question of mouse Ern2's role in goblet cell development and function, mucous layer integrity and the intestinal host-microbe interface.<br /> The emphasis, understandably, is on the mechanistic basis of the fitness-benefit arising from the duplication of the ancestral gene that gave rise to the two ERN paralogues: the broadly expressed ERN1 and mucous producing secretory epithelium-selective ERN2 of vertebrates.<br /> The finding that epithelial-selective depletion of mouse Xbp1, phenocopies the consequences of germline deletion of Ern2 is taken as evidence that Ern2 exerts at least some of its effects in goblet cells, via XBP1 - an effector common to the products of both ERN genes (IRE1a and IRE1b). This conclusion is plausible, but the data presented in figure 1A and 3A also suggests that the consequences of XBP1 depletion in terms of Alcian Blue positive cell number are 2X more severe than those of Ern2 deletion (~0.17 cells per crypt in wildtype, versus ~0.12 in the Ern2∆ and ~0.06 in the XBP1∆). This may arise from XBP1 having a broad role in tissue development, with secondary consequences on goblet cells, but may also arise from a contribution of IRE1a (the product of the Ern1 gene) to XBP1 splicing in goblet cells - in other words reflecting functional redundancy between the two ERN isoforms in mucous producing cells. The latter possibility seems at odds with the emphasis in the abstracts on non-redundancy of the two isoforms, which suggests important qualitative differences between the two, whereas the two genes may in fact be pullin.g in the same direction, in so far as their effector functions are concerned.

On 2021-07-29 20:22:05, user umueller wrote:

Box 1, left column, last paragraph: S is the selection differential, not the selection coefficient

On 2021-07-29 16:12:46, user Andrew Kropinski wrote:

I deeply regret that the 12 sequences which I analyzed using megablast are either Escherichia/Shigella phages or are essentially identical to Escherichia coli genome sequences.

On 2021-07-29 10:14:35, user Michael Coleman wrote:

This is a really interesting article on a topic we tend to take for granted and then realise we (or at least I !) just hadn't thought about and certainly couldn't explain. Some mechanisms for microtubule polarity sorting in axons had been previously proposed but were recognised as being insufficient to fully explain the observations. Very nice original science with important implications for nervous system development, axon regeneration and neurodegenerative disease.

Summary of findings

Unlike dendrites, axons have microtubules that are almost all oriented with their growing (+) ends outwards. The mechanistic basis of this is not completely understood. Axonal microtubules are in a constant state of dynamic equilibrium, with their + ends growing but being subject to periodic ‘catastrophe’ that shortens them, either by dying back from their previously growing + ends or by severing them to create two 'daughter' microtubules, each with the potential for new growth. Unlike in other cell types, axonal microtubules are not attached to the centrosome but form a tiling array along the axon composed of individual microtubules from a few microns to over 100 microns in length (see work of Peter Baas and colleagues). Some kind of relationship between this dynamic equilibrium and selection of polarity appears likely but it has been unclear what that might be.

To understand the mechanism, Jakobs et al used live imaging of microtubules in Drosophila axons in culture, labelled with EB1-GFP, which marks the growing tip. They find that during early axon growth in culture, microtubules with their + ends oriented distally have a growth advantage over those in the opposite orientation, so that over time + end-out becomes the dominant orientation.

First, they show that each microtubule growth events is (on average) longer if the microtubule is further distal in a growing axon and if the microtubule is oriented + end out. The difference between + end out, and in, microtubules is more marked distally.

Then, they measure the shrinkage distances in these same orientations and locations using double labelling of EB1 and tubulin. They use a mathematical model to show that + end-out oriented microtubules near growing tips have essentially unbounded growth (since the average growth event is longer in distance than the average shrinkage event), while in other locations and orientations average microtubule length stabilises because of the larger contribution of shrinkage events.

Using two methods to disrupt microtubule polymerisation (nocodazole and increased osmolarity) they then confirmed the importance of this +/- growth difference in establishing unipolarity. They also hypothesised that microtubule growth promoting proteins locally synthesised at the axon tip, such as p150, would explain the longer growth cycles of +end out oriented microtubules there, and supported this hypothesis with p150 knockdown and dominant negative mutants. Again, removing the growth length differential also removed the orientation difference.

Finally, they address the orientation imbalance in more proximal axon regions that is less easy to explain based on a p150 gradient. They propose a model in which dynein-mediated sliding of – end out orientated microtubules towards the cell body, and templating of new microtubules, essentially matching existing orientation bias, could explain these differences. No additional data are presented for this part but it clearly forms a new hypothesis for further testing.

Implications

Axonal transport deficits are an important driver of axon loss and neurological deficits. For example, mutations in the anterograde motor protein KIF5A are associated with hereditary spastic paraplegia, Charcot-Marie Tooth disease and ALS, all disorders of long axon degeneration in which distal regions are affected first. Toxic blockade of axonal transport, for example in vincristine neuropathy, is also an important cause of axon damage. This article sheds light on the basic mechanisms that establish, and presumably also maintain effective, directional axonal transport.

Severe defects in this process of selection would be expected to result in failure of neuronal differentiation or axon growth. The likely phenotypic outcome of a severe defect would be embryonic lethality but partial defects could also occur and could therefore underlie disorders of axonal transport even if axons do initially form and carry out the process. Indeed, p150 mutations are associated with ALS. It would be really interesting to know how such mutations affect microtubule polarity and whether this underlies pathogenesis in these cases of ALS, or indeed in any other neurodegenerative disorders. It is challenging to address this in vivo, even in animal models, because of the requirement for live imaging of microtubule growth so I am not aware of any previous studies, but it is in principle an achievable aim now this mechanism has been identified.

Limitations

At present these findings are limited to Drosophila axons (seemingly dispersed starting from the entire CNS?) so it remains to be confirmed whether there are similar patterns in mammalian axons, and in different neuronal subtypes (e.g., CNS/PNS, motor/sensory, etc).

Minor suggestions for improvement

Just a presentational thing but in Fig 1E legend, would it be clearer to say ‘blue, right to left downwards’ than ‘blue, left to right upwards’ since these microtubule are in fact growing from right to left? Or probably the colour-coding explained in part D is already sufficient without this extra explanation?

A bit more introduction to what is templating and sliding would be helpful.

It would be just marginally easier to follow without the switch in axon orientation between Figs 1-3 and Fig 4. But this is a minor point that perhaps just keeps our reversal learning sharp anyway!

Questions for the authors

Superficially, it could be imagined that the more stable an axonal microtubule the better, since they are so crucial for axonal transport. Yet, this is clearly not the case, otherwise the state of dynamic equilibrium would not have evolved. Does this new model for selection of orientation shed any light on what that advantage of the dynamic equilibrium is?

Studies of shrinkage events are so far limited to shrinkage from the distal end. Is there any contribution also from severing and how could that be measured?

If +end-out microtubules at the distal end have unbounded growth what eventually stops them? Something must do this in the end because otherwise a mature axon would be clogged with lots of microtubules extending right up to the distal tip. Is this one of the functions of severing?

In Fig 3b and c, there seems to be not only a decrease in + end-out growth distances but an increase in the growth of – end out microtubules. The same is true in Fig 3j and k when p150 is disrupted. Are these consistent observations and what could explain them? It would seem more likely that these interventions would disrupt microtubule growth regardless of orientation?

To what extent do you think similar mechanisms may operate in mature axons, or is this phenomenon limited to axon growth stages? At the very least it seems likely that they also recur during axon regeneratio but in this context it would be very interesting to know if there are CNS/PNS differences in vertebrates given the difference in axon regeneration.

On 2021-07-28 23:00:52, user Charles Warden wrote:

Hi,

Thank you very much for posting this preprint.

1) I mostly have experience with evaluating long-read assemblies on some BACs containing Eukaryotic sequence (say, between a half dozen and a dozen samples/assemblies). I think those also had duplications and repeats that added more complication than your average assembly.

While I think the most important message was a need to critically assess each sample's assembly on a case-by-case basis (without locking down one single assembly method and parameters in advance), I thought Canu tended to perform better than Flye on the samples that I have worked with.

So, I was surprised that I didn't see any Canu results in this preprint, even though I also noticed that you said you did a previous benchmark.

Since the title of the paper says "Bacterial," I guess checking for circular BAC constructs is slightly off topic. However, Canu does have a suggestCircular output column, and I found it useful for providing a good "initial" assembly for circular sequences (even if that particular column wasn't perfect).

It looks like you encourage the use of additional polishing, which matches my experience. I also strongly agree with the point "Trycycler is not a fully automated pipeline – it requires human judgement and intervention.".

However, is it out of the question to have comparisons with Canu assemblies, and/or have something to add for others like myself that found Canu to be preferable to Flye for other circular assemblies (with additional downstream steps, such as the MUMmer visualization recommended on the Canu FAQ)?

2) I encountered an issue clicking the Data availability link in the abstract.

I can find this repository (and I can also access the 2nd link):

https://github.com/rrwick/T...

Is that what you intended, or is there yet another link?

Thanks again!

Sincerely,<br /> Charles

On 2021-07-28 21:58:46, user Charles Warden wrote:

Hi,

Thank you very much positing this preprint.

I apologize if I overlooked something, but is your strategy doing something to try and gauge whether the "dark matter" is related to the COI gene or is off-target sequence that may be contamination?

As one example, I think there have been periods of time where top BLAST hits for PhiX sequence have said something else (because the lab didn't realize the PhiX spike-in could barcode hop into samples): https://www.biostars.org/p/...

Or, even if the PhiX sequences are already filtered, do you think DARN can help with things like this post-publication review thread:

https://www.nature.com/arti...

In that second example, you can do things like filter for sequences present in at least 2 reads in at least 1 sample. However, you can also see some troubleshooting that I think might be somewhat more similar here, and I am wondering if you think DARN might help identify groups of samples with a higher fraction of sequences that might not be a good representation the original biological sample.

Thank you again! I think the topic is quite interesting and important.

Sincerely,<br /> Charles

On 2021-07-28 15:43:12, user Alex Cope wrote:

Currently published at PNAS: https://doi.org/10.1073/pna...

On 2021-07-28 14:33:17, user CircaReader wrote:

Interesting paper with exciting results. There are certainly valuable implications from these findings. After reading through, I had several comments that I thought I would share.

1. Add legend to figure 1 x axis, “Days”. Perhaps p-day would be useful.
2. Is it standard to have text labels describing the results as seen in figure 2?
3. What is the purpose of the colored arrows in figure 2?
4. What is meant by ‘significant trend’ in figure 2/4? What test was used to determine this? Probably will need to clarify this statement

On 2021-07-28 10:18:32, user Mattia Deluigi wrote:

Very exciting contribution to the field, congratulations.

I have 1–2 suggestions regarding the structural comparison with the related neurotensin receptor 1 (NTSR1), e.g. as shown in Extended Data Fig. 8f. As inactive-state NTSR1 structure, the complex of rat NTSR1 with the full agonist neurotensin 8–13 fragment is shown (PDB ID: 4BUO). While this structure was obtained without G protein and adopts a closed conformation at the cytosolic side, the extracellular receptor half is virtually identical to other agonist-bound, active-state (or active-like) NTSR1 structures (e.g., PDB IDs: 6OS9, 4XEE, or 4GRV).

We have recently determined three structures of rNTSR1 bound to two inverse agonists. Compared to the agonist-bound structures, the inverse agonists caused a pronounced outward tilting of the extracellular ends of TM6 and especially TM7, and of ECL3. It may be interesting to point out the similarity between GHSR and NTSR1 in the large extracellular displacement of TM7 upon activation / inactivation.

PDB IDs: 6Z4S (structure with the least engineering), 6ZIN, 6Z4Q.

Full story DOI: 10.1126/sciadv.abe5504

We also shed additional light on the activation mechanism of NTSR1, and on how nonpeptidic agonists and inverse agonists modulate receptor activity. There are striking analogies and some interesting differences between our findings with NTSR1 and those on GHSR presented here (e.g., the involvement of residues 3.33 and 6.55; the importance of the aromatic cluster composed by residues 6.44, 6.48, 6.51, 6.58, and 7.42; cavity II in GHSR and a corresponding hydrophobic sub-pocket in NTSR1 crucial for activation / inactivation).

Presenting a slightly more detailed comparison of NTSR1 and GHSR may substantially expand our mechanistic understanding of the ghrelin family receptors and thus represent another valuable contribution to the field.

Kind regards and all the best,

Mattia Deluigi

On 2021-07-28 09:42:07, user Mattia Deluigi wrote:

Impressive work, congratulations.

Interestingly, the involvement of ECL3 in the modulation of CCKAR activation resembles what we have observed for the neurotensin receptor 1, for which we have recently solved structures in complex with agonists and inverse agonists. DOI: 10.1126/sciadv.abe5504

In NTSR1, agonist binding stabilizes an inward displacement of ECL3, coupled to an inward movement of the extracellular ends of TM6 and TM7.

Together, these findings may help design new potent synthetic agonists or inhibitors for peptidergic GPCRs.

All the best,<br /> Mattia

On 2021-07-27 21:49:16, user Mugdha Sathe wrote:

Note to readers<br /> The review presented here is combined from the submission of four undergraduate seniors from UW Seattle along with two additional colleagues. Each undergraduate submitted their solo peer-review report as part of their final project homework submission. <br /> Names: Noam Hezroni (noamh@live.com), Gianni L Landby (glandby@gmail.com), Michael Js Park (michaelparkjs@gmail.com), & MacKenzie Utley (mackenzietutley@gmail.com)<br /> Instructors: Dr Mugdha Sathe (mugdhas@uw.edu) & Dr Rebecca M Price (beccap@uw.edu)

Introduction<br /> The paper “Microfluidic guillotine reveals multiple timescales and mechanical modes of wound response in Stentor coeruleus” by Kevin Zhang, Lucas Blauch (https://doi.org/10.1101/202... posted Nov 12 2020), and colleagues is about quantifying and characterizing the single-cell wound response in Stentor coeruleus, a single-celled free-living ciliate protozoan. While wound repair in multicellular organisms is well-studied, wound repair in single cells is not as well understood. To address this, the authors used a microfluidic guillotine in order to inflict wounds with both low (Regime 1) and high (Regime 2) viscous stress to wound Stentor cells. They then used a fluorescence-based assay to visualize the timescales and mechanisms of wound repair in Stentor. The authors discovered three mechanical modes of wound repair: contraction, cytoplasm retrieval, and twisting/pulling. Furthermore, they concluded that Stentor requires 100-1000 seconds to close the wound and that the timescale depended on the severity of the wound. This manuscript’s experiments were well-designed with convincing results, and we recommend its publication after the major and minor points below are addressed.

Major Points<br /> 1. The authors observed that while most of the injured cells used at least one mechanical mode of wound repair, no cell used all three modes that they characterized. This raises the question about how/why a given mechanical mode of wound response is selected by a cell. For instance, cells wounded under Regime 1 preferentially use contraction or cytoplasm retrieval over the twisting/pulling method, but how does the cell choose to use one over the other or a combination of the two? Possible reasons include the energy cost of each method or potential correlations between cell sizes and the mode of wound response. Or the mechanical mode of wound response used could also be completely random.<br /> 2. In their conclusions, the authors address that the large variation in the time it took for cells to heal could be due to differences in initial cell size. They provide a supplemental figure for cell size variation (Additional File 1, Figure S4), but it would be useful to include such a figure within the paper itself. In doing so, the authors could compare the initial cell sizes with the time it took the cells to heal to determine whether cell size truly does correlate with the timescale of healing and whether it warrants further investigation.<br /> 3. Within the paper, the text states that cytoplasm retrieval was utilized by about 50% of cells wounded in Regime 1 and by only about 35% of cells wounded in Regime 2. Despite this, Figure 3B shows cytoplasm retrieval by cells wounded in Regime 2 rather than cells wounded in Regime 1. Referencing figures that align with the findings, for instance, representative images of a cell wounded in Regime 1 undergoing cytoplasm retrieval rather than a cell wounded in Regime 2, would better reflect the observations, and support the conclusions made in the paper.<br /> 4. The use of the microfluidic “guillotine” as intended was ensured through previous research utilizing and developing it, which is cited in this paper. The microfluidic “guillotine” could theoretically stop functioning and lead to no wounding of the cells, but the Sytox Green staining indicates that because the lack of any staining to a cell would suggest it was never wounded. However, it would be helpful to see a control for the microfluidic “guillotine” in which the cells travel through a similar apparatus but would not be cut. This would control for any contributions that travel through this device could have on wound healing.

Minor Points<br /> 1. Figure 1G indicates that cells wounded under Regime 2 took about 1000 seconds to heal, yet the brightfield and corresponding fluorescence images of wounded cells in Figure 1F jump from 630 seconds, in which a wounded cell with Sytox staining is still visible, to 7230 seconds. It would be easier to follow Figure 1F if it better aligned with Figure 1G and the observations made in the paper by including an imaging panel at or around 1000 seconds that shows a healed cell for Regime 2.<br /> 2. Within the text, the authors state that the cell’s oral apparatus was stained very brightly by immunofluorescence of acetylated tubulin in Figure 2. Yet Figure 2 itself does not have the oral apparatus labelled, nor is the oral apparatus mentioned in the figure legend. Per the text, the disruption or complete absence of the oral apparatus in cells wounded under Regime 1 or 2 indicated the severity of the wound and was useful in visualizing the cell’s healing process. Thus, it makes sense to label such an important visual aid in the figure itself, in addition to mentioning it in the text.<br /> 3. It would be easier to follow the contraction mode of wound response in Figure 3A in the paper if, instead of an arrow pointing to the wound, a red line was drawn through the wound to show the wound’s diameter and how it changes as time proceeds, as was included in Additional File 1, Figure S3.

On 2021-07-27 00:40:11, user Charles Warden wrote:

Hi,

Thank you very much for posting this preprint.

I can think of a number of questions, but I hope these are the most important:

1) I noticed that you filter for samples with “>10,000 and <500,000 classified microbial reads.”

Does this filter also apply to returning results to customers?

Based upon data / results I will mention in 2), I believe my cat’s samples may have had ~25% bacterial reads (at least for the ~15% WGS sample). For the earlier lcWGS samples (for “Health + Breed” kits), I thought they were 2-4 million reads each. If the Dental Health Test was similar, then I think my cat’s sample(s) would have had “<500,000 classified microbial reads”. However, I am not sure if I am misunderstanding something?

2) You can see the various measurements for my 1 cat here:

https://github.com/cwarden45/Bastu_Cat_Genome/tree/master/basepaws_Dental_Health_Test

My question is whether there is some figure in this preprint that I can use to compare the scores for my cat (or any customer can do for their cat), compared to the density distributions cases and controls used for the risk estimation in this paper (for what I might call a training set)?

I am also curious if those proportions say consistent among batches (and filtered samples or full samples, if they are all returned to customers).

I think this might be similar to Figure 3, but I am not sure how to map the scores for my cat.

3) Relative to the ordinary sample collection, is anything added to the liquid to limit or prevent post-collection bacterial blooms?

If something was already added, is there any evidence that helps restrict / prevent bacterial growth from earlier tests?

4) The Competing Interest Statement currently says “The authors have declared no competing interest”. However, because you are working for a company that offers a Dental Health Test based upon these results, I believe you need to declare a competing interest. Is there something that I might be missing?

Thanks again!

Sincerely,<br /> Charles

On 2021-07-25 23:20:14, user Luke Eberhart-Hertel wrote:

This pre-print is a major revision of a previous pre-print found here: https://www.biorxiv.org/con...

Please also see the comments related to the prior version for links to a pre-publication review of the work on Publons, and our open-access response to this review.

On 2021-07-24 18:06:14, user Jin Yu wrote:

published online: https://doi.org/10.1016/j.b...

On 2021-07-23 21:07:12, user Etsuro Ito wrote:

Dear my Lymnaea friends,<br /> Thank you for your wonderful job for the preparation of database of proteomics.<br /> We are now applying LC-MS to our Lymnaea studies and thus your database encourages to advance our studies.<br /> Please continue such useful work to not only Lymnaea researchers but also other molluscan researchers.<br /> Best,<br /> Etsu

On 2021-07-23 10:24:10, user Lore Pottie wrote:

Our preprint has been published (https://doi.org/10.1371/jou.... The link will follow forthcoming.

On 2021-07-22 23:32:19, user Jessie Lopez wrote:

There is a DIY audiobook of this paper available on YouTube if you would rather hear the author read the paper.

https://www.youtube.com/wat...

On 2021-07-22 09:52:07, user Isabella H wrote:

Dear Authors,<br /> I have some suggestions regarding your preprint.<br /> Although the investigation of eDNA from air is an understudied field let me make you aware of some studies which were mainly focused on microbial an fungal bioaerosols (10.1016/j.scitotenv.2020.144092, 10.1073/pnas.0811003106, 10.1093/femsec/fiy031) and collection methods (e.g. 10.1021/acs.est.7b01480, 10.1371/journal.pone.0141158, 10.1128/AEM.01589-18). Maybe you want to include some of this literature in your manuscript.<br /> In addition it would be great if you could add the primer sequence of Mam1 and Mam2 to the manuscript. Regarding figure 2, I guess it could help if you included boxplots in 2a and maybe a logarithmic scale for Fig.2b or you may think to exclude the two outliers form the visualization.<br /> All the best<br /> Isabella

On 2021-07-22 08:29:35, user Alizée Malnoë wrote:

The manuscript by Chazaux et al. presents the characterisation of ape1 mutant alleles of Chlamydomonas and shows that they do not acclimate as well to increases in light intensity as does the wild type, consistent with the Arabidopsis ape1 mutant described in 2003 (Walters et al.). The authors have clearly shown that APE1 is a thylakoid membrane protein required for the proper function of PSII at high light intensities. The absence of APE1 leads to changes in thylakoid architecture and an altered composition of photosynthetic complexes (20% less D1/D2 and 20% more LHCII in ape1). In this new submission, they have unequivocally demonstrated that APE1 directly interacts with PSII (co-migration of APE1 with the monomer and dimer of PSII which is altered in deltaPSII but not in deltaPSI as well as cross-linking with PSII). Furthermore, they have shown that in the absence of APE1, PSII supercomplexes (sc) accumulate (BN-2D and in vivo measurements of antenna size). This accumulation of PSII sc leads to photoinhibition and the authors propose that the role of APE1 is to modulate the formation of supercomplexes for optimal photosynthesis.

I largely agree with these conclusions and think that this body of work moves the field forward. I have several comments however that should be addressed (see below). My main issue lies in the interpretation of Figure 7A and B (photoinhibition experiment) and lines 314-318. My previous comment on that point still stands: if D1 damage would be increased in ape1, we should expect a lower amount of D1 upon inhibition of repair. However according to Fig7B, this is not the case (ape1 has apparently the same level of D1 as in wt). The mph1 mutant described by Liu and Last 2015 Plant Journal 82:731–743, is proposed to have more PSII damage which translates as less D1 accumulation upon inhibition of repair (coincidentally also in Figure 7B!) Granted this is in Arabidopsis and the timepoint for D1 is 2h HL, I would still recommend repeating that experiment e.g. a 2h HL+inh checking D1 accumulation. As is, what I take away from Fig7B is that the rate of repair is slower in ape1. If I understand correctly, the authors imply that because of the large antenna size, D1 is more prone to photodamage but given this other interpretation (slower D1 repair), how about the idea that repair is slower due to D1 being more trapped in PSII dimer and sc? You discuss this idea in paragraph from line 519.

This interpretation better fits the conservation of APE1 in cyanobacteria which do not possess PSII sc and agrees with the possible constitutive function in maintenance of PSII state described line 516. The title of the article should then be changed accordingly.

Major comments

Line 71 - I understand the decision from the authors to cite reviews and state that “many” proteins are involved in PSII biogenesis, maintenance and repair. I would however recommend citing Liu and Last 2019 Frontiers in Plant Science 10:975 and either introduce or discuss more what factors are known to modulate PSII sc (e.g. from this review MET1; in met1 mutant less PSII SC, accelerated D1 degradation; as stated above should we expect ape1, more PSII SC, slower D1 degradation?)<br /> Line 114, line 256 - RC47 should be defined as reaction center 47 (see Komenda J 2006 J Biol Chem 281: 1145–1151), it can be both an assembly or repair intermediate.<br /> Line 129 - “Atape1” should be written as “The Arabidopsis ape1 mutant” for clarity.<br /> Line 229 – “and is not strictly required for their assembly”, I do not agree with that statement, chaperone required for assembly can accumulate to wild type level in absence of the main subunit of a given complex.<br /> Line 319 to 325 – Description of partial complementation (lower chla/b in HL and intermediate GPX5 accumulation) and possible explanation (since APE1 protein accumulates to wild type level) should be provided.<br /> Line 341 – for this paragraph, it would help the non-initiated to add more introduction on alpha and beta centres and what is meant exactly by “PSII heterogeneity”.<br /> Line 573 – The method for generating the mutant library that led to isolation of ape1-1 is missing<br /> Line 676 - Provide amino acid number where the sequence used for recombinant proteins starts<br /> Fig1D,E – The data is the same as in the previous submitted version, ape1-1 should say drbcL ape1-1 in Fig1D,F and 1E control is cc-4533.<br /> Figure 2B – orientation of protein? please indicate stroma and lumen compartment<br /> Fig2C – comment on reason for signal to be lower in isolated thylakoid compared to total cell extracts, usually the opposite occurs, unless APE1 is dual-targeted?<br /> Fig4B – comment on more PSI(2) band in ape1 mutant<br /> Fig5 – comment on APE1 oligomers. I agree that cross-linking with PSII demonstrates that APE1 is not a free protein but upon solubilization different subcomplexes are resolved, especially in the deltaPSI and deltaPSII mutants.<br /> Fig7C – comment on nature of the band right below E and above F enriched in the complemented line.

Minor comments<br /> Line 147 and other instances – Figures are called twice<br /> Line 184 – should read “violaceus”<br /> Line 286 – change decorated to probed<br /> Line 401 – should read “photoinhibition”<br /> Fig6A – annotate nature of B5 and 8 to help reader<br /> Fig9 – is this in MIN? how many days? Light intensity?

Alizée Malnoë (Umeå University) – prompted by a journal; I’m an assistant professor, my research group investigates the molecular mechanisms of plant photoprotection. I performed a bachelor project internship in Gilles Peltier’s group, Jean Alric and Xenie Johnson are former colleagues from the group I performed my PhD studies in.

On 2021-07-22 08:24:50, user Alizée Malnoë wrote:

Nawrocki et al. with this manuscript make an important step forward towards understanding the molecular origin of nonphotochemical quenching (NPQ) qI using the microalga Chlamydomonas as their study system. Indeed upon high light stress, chlorophyll fluorescence quenching is observed attributed to PSII photoinactivation, and this work demonstrates that it stems from PSII reaction center and that degradation of D1 by FtsH is required to relax to an unquenched state. The authors further show that qI formation is more rapid in presence of oxygen but independent of PSII activity (in DCMU) and propose that qI is due to oxidative modification to chlorophyll molecules of PSII reaction center (RC). Accordingly a minimal model was built with qI-ON RC and qI-OFF FtsH-processed/broken RC which successfully fits the experimental data indicating that photoinactivation mechanisms at donor and acceptor sides co-occur.

Here are some suggestions/comments. Looking forward to discussion!

The first point pertains to semantics, I would suggest using the word “photoinactivation” of PSII instead of “photoinhibition” as much as possible. I was convinced of this idea by Barbara Demmig-Adams when I worked on a review about qH, reserving the term photoinhibition for decrease in CO2 fixation. We proposed a possible new definition for qI that would be quenching due to photoinactivation of D1 rather than due to photoinhibition, as qZ and qH are also photoinhibitory (in that they decrease CO2 fixation). Take a look here, see intro, section 1.1.2. and 4.: https://doi.org/10.1016/j.envexpbot.2018.05.005.

Title: I’d suggest a more descriptive title of findings stating where qI stems from, e.g. with oxygen sensitization of D1 as the origin of qI. As is, one could understand the title as qI doesn’t exist (dogma rose and fell kind of idea) but you mean instead molecular origin/mechanism of qI induction and relaxation, right?

Throughout text, I’d suggest to use the word “relaxation” instead of “quenching loss”. The word loss is used elsewhere to describe fluorescence decrease i.e. quenching and it can be confusing to have the word “loss” used for both quenching induction and relaxation (e.g. line 94 loss= relaxation; line 96 loss= decrease).

I’d also suggest for clarity to use “new synthesis” instead of “repair” (e.g. line 101), as repair encompasses both degradation and new synthesis, it seems confusing to read that qI relaxation is independent of repair but relies on degradation.<br /> e.g. Figure 2 title becomes: qI is transient and its relaxation is independent of new D1 synthesis<br /> and Figure 3 titles becomes: qI relaxation is due to PSII core proteolysis by FtsH

Also for clarity, in title of Figure 1 add - is a quenching “due to energy dissipation”-<br /> And line 452 - Using quenching -add “of Fm” - might be beneficial for (f)uture studies

Line 202 FtsH-mediated (name of protein uppercase here; not mutant italic lowercase)<br /> Line 211 side -> sites

There’s a lot of crucially ;-)

To go further in the discussion, here are some points that could be interesting to raise:<br /> - Addition of lincomycin blocks synthesis of all chloroplast-encoded proteins, impact on qI formation/relaxation.<br /> - qI relaxation in presence of nuclear gene synthesis inhibitor.<br /> - Slower relaxation of qI in the dark compared to low light (at least after 30min HL).<br /> - qI transient: explanation for differences between strains (e.g. 1009 vs. 124).<br /> - ftsh complemented line (in ftsh1-1), comment whether less qI compared to control because more repair enabled; due to higher level of FtsH in that complemented line compared to one in ftsh1-3?<br /> - formation of qI site precedes cleavage (line 242) would need deg mutant to definitely say that. Might be better to say that it precedes D1 degradation (at timepoint 0 there are some D1 fragments, so there has been cleavage already before HL starts).<br /> - damage at acceptor side triggers cleavage in the lumen? (1995 Plant Phys D1 qI https://www.jstor.org/stable/4276408. After HL stress, decrease in D1 but DCMU-binding sites 2x higher vs. D1 detection by antibody; propose preferential cleavage in the lumen).<br /> - if oxygen sensitization proceeds by PSII charge recombination (line 404), then should qI be enhanced in DCMU? Compare with hydroxylamine (HA)+DCMU to test it.

• These publications about Arabidopsis mutants seem consistent with your findings, what do you think (although hard to say for sure without Fm quenching data):<br /> 2014 Plant J Kato & Sakamoto, https://onlinelibrary.wiley.com/doi/full/10.1111/tpj.12562<br /> Fig5b,c: similar levels of D1 as control (fug1) in fug1 var2 stn8 (~ equivalent Chlamy would be ftsh stl1) but with lower Fv/Fm so D1 is removed in fug1 var2 stn8 but would be quenching more than in fug1 due to more ROS in stn8 background; fug1 var2 has most damaged D1 so most quencher (OR alternative hypothesis phosphorylated RC are quenchers?)

2011 Plant Cell LQY1, www.plantcell.org/cgi/doi/10.1105/tpc.111.085456 <br /> 2014 Plant Cell HHL1 LQY1, www.plantcell.org/cgi/doi/10.1105/tpc.113.122424 <br /> lqy1, hhl1 have faster rate of degradation and low Fv/Fm after HL due to higher Fo but Fm stays the same; that would be consistent with faster degradation, less quenching of Fm.

2017 PNAS MPH2 www.pnas.org/cgi/doi/10.1073/pnas.1712206114 <br /> Impaired degradation of D1, low Fv/Fm, more quenching of Fm.

Schematic model: the different shades of grey in the wheel do not represent light/dark, correct? maybe would be clearer to make it according to light treatment. The dash-line is not described: is it the alternative hypothesis to cleavage, that full degradation is required to relax qI? I like the purple scribble on Nter of D1 to signify ‘hey, degrade me!” ;-)

Alizée Malnoë (Umeå University) – not prompted by a journal; I’m an assistant professor, my research group investigates the molecular mechanisms of plant photoprotection. Catherine de Vitry was my PhD studies advisor.

On 2021-07-22 04:11:25, user Alex Crits-Christoph wrote:

Quickly after the publication of this preprint, it appears the data <br /> from Wang et al. 2021 was re-uploaded to a public database by the <br /> original authors here:

https://ngdc.cncb.ac.cn/bio...

On 2021-06-29 18:48:36, user Jesse Bloom wrote:

This version of the manuscript (v2, https://www.biorxiv.org/con... is a revision to the original preprint (v1, https://www.biorxiv.org/con....

A detailed description of the revisions and the rationale behind them is available here (http://disq.us/p/2hwapg6) "http://disq.us/p/2hwapg6)") as a comment on the first version of the manuscript.

On 2021-06-28 03:16:33, user Jesse Bloom wrote:

I thank Stephen Goldstein and an anonymous commenter with the username ACC for their helpful comments above on the initial version of the bioRxiv pre-print.

Below I outline how I have addressed these comments in a revised version of the manuscript.

These revisions distinguish version 1 of the bioRxiv pre-print (uploaded on June-18-2021) and version 2 of the bioRxiv pre-print (uploaded on June-27-2021). To see the details of all revisions, go to https://github.com/jbloom/SARS-CoV-2_PRJNA612766/compare/initial_bioRxiv_version...second_bioRxiv_version#diff-983b58b8186b0a4ed7f280f258cdab3eb0dd7d5136f8ac361ba982a43cfb7136 and if you are specifically interested in the manuscript (rather than the code) find the file called paper/paper.tex and click on Load diff to see all the changes.

Below I summarize the revisions:

ADDING E-MAILS REQUESTING SRA DELETIONS<br /> Both commenters pointed out that it is hard to be sure why the data were deleted from the SRA. In the revised version, I have added additional information relevant to this point.

Specifically, after I e-mailed the pre-print to the NIH on June 18, 2020, they sent me a copy of the e-mails from Wuhan University requesting that deletion. I now include these e-mails (with the redactions and highlighting made by the NIH) as Figure 6 of the revised manuscript (they are also at https://raw.githubusercontent.com/jbloom/SARS-CoV-2_PRJNA612766/main/paper/figures/SRA_email.png). I have added these e-mails instead of referring to the NIH statement (as suggested by Stephen Goldstein) because these e-mails contain more detail than the NIH statement.

In the revised version, I have refrained from assigning a motive to the authors. However, I do point out that although their e-mail said they were submitting the data to another website, I can find no website with the data. I also point out that whatever the motive, the practical consequence of removing the data from the SRA was that no one was aware of its existence even if it was technically available both on the Google Cloud and in a table in journal Small. In particular, removing the sequencing data from the SRA meant it was not in the list of locations from which the joint WHO-China report collected their data.

While I think the commenters were correct in asking me to moderate my strong suggestions about the motives for removing the data, I similarly think commenter ACC is incorrect when (s)he says that the fact that the paper corresponding to the data was published shows with high confidence that the authors were not trying to obscure the data. In fact, the authors had uploaded a pre-print to medRxiv on their study in early March, and there is no mechanism for deleting a pre-print (unlike for SRA data). So the authors were committed to having the manuscript permanently available as soon as the pre-print posted in early March.

EXAMPLE SRA DELETION FROM ANOTHER STUDY<br /> In the original pre-print, I included as Figure 2 an example of how SRA data are removed by showing an e-mail illustrating this process for data from a different study on pangolin coronaviruses. This was an e-mail excerpted from page 50 of https://usrtk.org/wp-content/uploads/2020/12/NCBI-Emails.pdf. I included the e-mail because it was the only publicly available example I could find of the SRA deletion process.

However, both commenters suggested this e-mail was confusing because it was from a different study. I accept this point. In addition, as described above, I subsequently received from the NCBI the actual e-mail to withdraw the BioProject relevant to the current study, so have included that as a more relevant example of the deletion process.

The commenter ACC further suggests that my original reference to the deleted pangolin coronavirus accessions SRR11119760 and SRR11119761 is a “mistake” because those data are available on the SRA again. I checked, and this is indeed the case, but those two SRA files only re-appeared on June-18-2021 (according to timestamps from vdb-dump --info, see: https://github.com/jbloom/SARS-CoV-2_PRJNA612766/blob/main/paper/figures/SRR1119760_SRR1119761_obj_timestamps.png) which is the same day that I submitted my pre-print to bioRxiv. I obviously had no way of knowing these two pieces of missing data would reappear on the SRA after a 15-month hiatus almost concurrently with submission of my pre-print.

MAKING CLEAR DATA ARE STILL LISTED IN TABLE OF PAPER IN SMALL<br /> Both commenters correctly pointed out that even after the SRA deletions, the mutations were still available in a table in the paper in the journal Small. In the revised version, I have added additional text to make that point very explicitly (it was mentioned before, but only in passing). However, I also note that Small is primarily a chemistry journal that is not read by virologists, and the practical consequence of removing the sequencing data from the SRA is that no one was aware of the list of mutations in the paper in Small until I recovered the sequencing data from the SRA.

HOW TO REFER TO EARLY EPIDEMIC SEQUENCES<br /> The commenter ACC says it is too vague to describe the samples as from “early in the epidemic.” However, this is an exact quote from the Wang et al pre-print, where they describe the samples as being from outpatients “early in the epidemic.” The final published version of Wang et al changes this to “early in the epidemic (January 2020),” but neither the pre-print or paper give more exact dates of sample collection. I therefore have retained this description since it is how the authors of the study themselves describe their samples.

IMPORTANCE OF THE NEW SEQUENCES<br /> Both commenters correctly emphasize that the new sequences do not transform our understanding of viruses in Wuhan at this time. Rather, they provide modest additional information on viral diversity at this time, including more evidence for the C29095T mutation (which makes the virus more similar to the bat outgroup viruses). I fully agree that the data are not transformative, but contend that any new data are valuable. The original manuscript already made this point: for instance, clearly saying that the data supported the prior inferences of Kumar et al about proCoV2 (which was even used as the alignment reference). However, in the revised version I have further emphasized the data are supportive of existing conclusions by Kumar et al and several others, while leading to incremental advances (such as possible importance of C29095T). However, I am unsure how to respond to commenter AAC’s criticism of the title and abstract: the title says “sheds more light” which is a modestly worded phrase that is appropriate for the moderate increase in knowledge that has accrued from these new sequences. These are the only new early Wuhan sequences we have had in over a year, so I consider that a net plus for scientific knowledge even if it’s not transformative.

DISCUSSION OF HUANAN SEAFOOD MARKET<br /> In the original manuscript, I had a paragraph describing why I think two of the theories about the difficulty of rooting the SARS-CoV-2 are rather unlikely: the idea that RaTG13 is faked to confuse root placement, and the idea that there were multiple zoonoses from multiple markets. Both commenters raise various questions about this part of the manuscript. I still think both of these theories are unlikely, but discussing them in detail is not central to the main points of the paper. So I have shortened to just mention that I think that the two-market zoonosis is less plausible for the reason explained by Trevor Bedford here https://twitter.com/trvrb/status/1408080716286414852.

ADDITION OF SRR11313490 AND SRR11313499<br /> In the original manuscript, I could recover all of the deleted sequence data from the Google Cloud except SRR11313490 and SRR11313499. After I posted the original pre-print, I was contacted by several individuals who realized that they had downloaded copies of those data prior to the data deletion in June of 2020. I have added those new data in the revised manuscript. It turns out that their addition makes no meaningful difference, as both are from low-coverage samples for which it is still impossible to call meaningful sequence information.

PROPER TRIMMING OF THE READS<br /> It was brought to my attention by Brendan Larsen that I had failed to properly trim the sequencing data in the original analysis, and so was analyzing the primer overlap sites as well as the sequenced region. I have masked these sites in the analysis in the revised manuscript. This also makes no meaningful difference in the results, since no mutations of interest are in these binding sites, although it does slightly reduce the fractional coverage over the region by opening small gaps at the primer binding sites.

On 2021-06-28 02:48:27, user Stephen Goldstein wrote:

1. I think including an unrelated email to the SRA was unwise. It’s a reasonable inference from this that Chinese scientists somewhat broadly are involved in unscrupulous data handling and sharing practices. My understanding from others with respect to that specific email is that the data in question is back on the SRA, and the pangolin CoV sequences associated with that paper are available on GISAID. Implicating researchers unrelated to the Wang et. al. paper in this matter seems unfair. I don't think it serves a positive purpose but can have a negative connotation for Chinese researchers.

2. It's of course true you recovered the raw data files and you do reference Wang et al preprint and paper. However, I think you need to acknowledge that Wang et. al. specifically describes the mutations assigning these sequences to lineages A and B and even reference the lineage split (called L and S at the time). So while the raw sequences are newly recovered, the key information gleaned from them was not concealed. Your response on twitter that the data are less useful for analysis purposes in a paper table is something you can bring up to still support your argument that this was underhanded (though I disagree about the strength of evidence for this). But I think currently the reader comes away thinking not only the raw data but the genetic diversity information associated with it was concealed and as you know, this is not the case.

3. In general, it doesn't surprise me at all that the earliest sequences recovered might not actually represent the first infections. Since the outbreak didn't really catch attention until super-spreading at the Huanan market, almost all viruses preceding that went un-sampled. Uf the first human infections were in November as calculated (maybe at Huanan, maybe not, maybe there and somewhere else) then these viruses could not be the first sequenced examples and in fact none of the first sequences likely exist. So I don't think the discordance between the first reported sequences being more distant from the bat viruses is unusual, even if Lineage B is derived. I would argue it's actually expected. It may be particularly difficult to identify the first cases of a respiratory disease, often with unremarkable symptoms, then infections with a more unusual presentation.

4. I agree A may be a better root than B, though the proper route may also be between them. However, the details of this particular rooting issue is somewhat beyond my phylogenetic expertise.

5. It does not necessarily follow, however, that B is descended from A in humans. I think it's just as likely (or more for the below reasons) this split occurred in an intermediate host and represent independent spillovers. These sequences are from January, WA-1 is from January, there's one A virus from Dec (maybe?) in the WHO report. The existing evidence is therefore consistent with contemporaneous introduction of these lineages, rather than lineage A entering the human population first and B diverging from within lineage A diversity. Apparent intermediate sequences may result from early Illumina pipelines calling low coverage bases as Wuhan-1 (the reference) making it appear that some LinA sequences were LinA+a B mutation, though this requires additional study. There is precedent for diversity of SARSr-CoVs arising in an intermediate animal reservoir. Among four animal sequences of SARS-CoV sampled in spring 2003, they differed by 0 to 8 nucleotides in the spike gene, following several months of transmission among animals in wildlife markets, which were not shut down until the following winter.

6. Given the above, the Huanan market, if it was a spillover site, is certainly not the only spillover site. The Lineage A virus in the WHO report was linked to an unnamed market and one beneficial outcome of your work highlighting these sequences would be if epidemiological data can be linked to these sequences. I believe Huanan is a plausible spillover site with subsequent human-to-human transmission for Lineage B. The limited infections in early December (and molecular clock analyses) point to perhaps a mid-late Nov introduction there with limited onward transmission for some time before super-spreading commenced.

7. In terms of tone, I suggest sticking to the findings and staying away from assigning motive, in particular to individual researchers in undoubtedly difficult circumstances. The Chinese government has obviously been obfuscatory throughout this pandemic as with most things. Notably, the most well-documented obfuscation related to early stages of the epidemic was the denial to the WHO team that live mammals were sold at Huanan, which we now know to be untrue. Criticism of the Chinese government is therefore firmly within bounds. Based on the limited information available, I believe extreme caution with respect to criticizing the Wang et. al. authors is warranted.

8. You obviously need to add something in response to the NIH statement about the data removal, and the revelation that eight other data sets were also removed from the SRA.

-Stephen Goldstein, PhD

On 2021-07-21 14:37:48, user Taras Oleksyk wrote:

A modified and improved <br /> version of this article has been published in GigaScience. doi: 10.1093/gigascience/giaa159

On 2021-07-21 02:01:35, user ygc_smu wrote:

now published on The Innovation, https://doi.org/10.1016/j.xinn.2021.100141

On 2021-07-20 14:01:52, user Mel Symeonides wrote:

(My peer review no longer appears because it was made on the first version of the article, and it unfortunately did not receive a response before the manuscript was updated. I am thus reposting it here on the second version, hoping for a response. The review is identical to the first one, apart from the comment on the new sequencing data.)

With this study, Patterson et al. present a potentially very significant finding: that SARS-CoV-2 antigen persists in non-classical monocytes from Long COVID patients up to 15 months after the initial infection. The data supporting this finding are of moderate to low strength, as presented, primarily due to a wide range of major and minor presentation issues that are listed below. Most of these can be addressed easily, though it is unclear if some additional controls may be required. Finally, some orthogonal approaches are suggested that could be potentially very valuable in terms of increasing confidence in the findings (namely microscopy and immunoblotting), though these are not essential for the interpretation of the results as shown.

The authors are to be commended for tackling Long COVID head-on and getting right to the heart of the matter in terms of finding the pathological cause of this disease. That said, unfortunately, this manuscript requires considerable revision in order to be interpretable and allow others to reproduce the findings (which will be of critical importance, given their potential significance).

Major issues:

• Table 1 and the accompanying text seems to indicate that PBMCs were tested for the presence of viral RNA by ddPCR. However, in the Material/Methods section, it is stated that nucleic acids were extracted from plasma, not from PBMCs. Please clarify this point as it is of critical relevance. Indeed, both plasma and PBMCs should have been individually tested in order to determine whether viral RNA was solely intracellular.

• It is very unclear what Figure 2 is presenting. Presumably each row represents a different subject, but it is not denoted which subject belongs to which group, making intepretation very difficult. I presume this was an ommission.

• Supplementary Table 1 was not provided, making it very difficult to evaluate the flow cytometry data. Even if that table were present, the methods provided for flow cytometry are very sparse. What steps were undertaken to establish the specificity of the Spike antibody? Was the Spike staining done after fixation and permeabilization? Was PE conjugation of this antibody done in-house, and if so, using which kit, and how was it verified that the conjugation and quenching were successful and that staining was specific within the context of the entire antibody panel? Were FMO controls done in the context of this new panel that includes the S1 antibody? Was Fc block included? etc.

• The newly-added sequencing data are difficult to interpret. It seems that the authors interpret the poor sequence coverage as indicative of non-replicating virus and in line with high Ct values, yet they do not seem to comment on the fact that there is nevertheless some seemingly full-length viral genome present in these cells! This is a potentially very important finding and its source will need to be investigated. Additionally, the sequencing results are inconsistent with the PBMC RT-PCR results, where only LH5 was positive, yet LH1-5 all had similar sequence coverage. The names of the samples in Table 2 do not correspond to any other name in this manuscript, clearly they were not renamed as they should have been. In fact, some of these sample names look curiously like name initials, which is a potential study subject data privacy issue. Finally, it does not seem that any healthy controls or previous COVID/non-LH subjects were tested in the same manner, which would be very valuable information.

• In general, the Figure Legends are very sparse and should be much more descriptive.

Minor/moderate issues:

• Table 1 shows that one of the study subjects was asymptomatic. Where is this subject grouped in the subsequent analysis? ALso, "NS" is not defined, presumably it means "nasopharyngeal swab"?

• In Figure 2, left column, the CD14/CD16 gates shown were not applied equally from sample to sample. Furthermore, in the middle column it looks like S1+ non-classical cells tend to have a low-SSC profile, while S1- cells have a high-SSC profile that clusters together with intermediate cells. This suggests that the intermediate/non-classical discriminating gates may not have been set appropriately.

• The quantification shown in the middle column of Figure 2 is labeled "CD16+CD14+COVIDS1+", however no "CD16+CD14+" subset is defined. Presumably the authors refer to the aggregate of the "CD14++CD16+" intermediate and "CD14loCD16+" non-classical subsets. This should be clearly stated as it makes interpretation of the data shown very difficult. Additionally, the quantification is based on the aggregate population, whereas based on the color coding, one would expect individual quantification for each subset. Given the relatively very minor contribution of the intermediate subset to the observed Spike S1 signal, it is unclear why this was included at all in this plot - why not just show the non-classical subset and base the quantification solely based on that, or alternatively, show quantification of each subset rather than their aggregate?

• The labeling in Figure 3 could be better, the angled X axis labels are very difficult to follow. Maybe just indicate the monocyte subset as a title above each plot, and/or label each plot as a subfigure?

• No information is provided on the statistical analyses done.

• I did not look into all the cited work, but in one case (ref. no. 19) was puzzled to see that a review article was cited in which the relevant information was in turn derived from a single primary research article. Surely it makes more sense to just cite that primary research paper rather than the review?

General comments:

• Why was S1 the only SARS-CoV-2 antigen stained for? One would expect that you would have quickly tried to look for other viral antigens, particularly Nucleocapsid, in order to begin to understand whether there might be virus particles present, especially since you found viral RNA in some samples. Additionally, some microscopy data on sorted non-classical monocytes would have been very valuable to validate what you see by flow cytometry, also because one could then evaluate whether the Spike signal in these cells looks like the expected pattern for protein being actively synthesized by the cell and present on the cell surface, or whether it is captured antigen from some site of viral persistence and is sequestered in some intracellular compartment. Finally, a Western blot for Spike (and other viral antigens) in flow-sorted monocytes would be of immense value to further validate the presence of this antigen and observe the state of the protein - indeed, it is rather odd that you seemingly went for LCMS before trying either microscopy or a Western blot!

• The potential connections with the CX3CL1 pathway mentioned in Discussion are very interesting. Unfortunately, the authors have not demonstrated any elevation of CX3CL1 associated with severe acute COVID or long COVID disease, nor the presence of CX3CR1 on the particular cells of interest. If such data exist, please present them, otherwise this Discussion is rather speculative and much more work will be required to frame it in the appropriate context for a primary research paper. Alternatively, this discussion might be better suited for a separate Review article.

• Much of the published work on Long COVID and other post-COVID conditions such as MIS-C is omitted here, and should be cited and discussed as appropriate.

Mel Symeonides, Ph.D.<br /> Postdoctoral Associate<br /> Department of Microbiology & Molecular Genetics<br /> University of Vermont<br /> Burlington, VT

On 2021-07-20 13:19:09, user Ankush Sharma wrote:

The typographical error in Github repository link . Correct link for accessing the repository is https://Github.com/Eskeland...

On 2021-07-20 08:45:57, user Dr. Tyeen C Taylor wrote:

This article is now published in Frontiers in Forests and Global Change at: https://doi.org/10.3389/ffg...

On 2021-07-19 20:54:30, user stephens999 wrote:

A Review of Zheng et al, Universal prediction of cell cycle position using transfer learning, by Matthew Stephens

This paper provides a new approach (tricycle) for predicting the<br /> position of a cell in the cell cycle. The approach claims to work<br /> regardless of cell type, species and sequencing assay.<br /> There are several things to like about the paper. In particular,<br /> the tricycle method is very<br /> simple: i) compute the first two PCs on<br /> 500 annotated cell-cycle genes in a data set where cell cycle<br /> is the primary source of variation; ii) project<br /> any future observations to this 2-d embedding and compute<br /> the polar angle to predict its cell cycle<br /> position. Further, the empirical results are promising.<br /> At the same time I think the paper<br /> could be substantially improved by removing or<br /> reducing some of the less innovative parts, toning down some of the rhetoric,<br /> and focussing on the most convincing empirical results. My comments expand<br /> on these suggestions.

Main comments:

1. I found most of the material on PCA not to be<br /> especially novel or interesting. The use of PCA to determine cell cycle<br /> position has a long history (including many papers cited here),<br /> and existing mathematical results already go far beyond<br /> the analysis presented here. The behavior of PCA on cyclic phenomena<br /> is much more general than presented here, and does not rely on sinusoidal<br /> functions or "two distinct peaks" etc. Rather it stems<br /> from the result that cyclic phenomona lead to circulant covariance matrices,<br /> and all circulant matrices have the same eigenvectors:<br /> the columns of the discrete Fourier transform matrix. The result<br /> is that, when the covariance patterns primarily reflect cyclic phenomoena,<br /> the first two PCs will form a circle/ellipse.<br /> See Novembre and Stephens (2008) and references therein for further discussion.<br /> Figure 1 is useful for summarizing the method, but most of the other<br /> material could be condensed or removed and I think the paper would be improved because<br /> it would better focus on what is actually new and interesting, the tricycle<br /> method (currently not introduced until p6) and the empirical assessments of its performance.

2. The paper left me asking myself this: what is the strongest empirical support that tricycle cell<br /> cycle assignments work in practice? To me, Fig 5 panels c and g are the most convincing, because they are quantitative<br /> comparisons with an alternative technology (and one that is often considered the<br /> "gold standard" in this area). I also liked the quantitative comparisons with other<br /> methods, and it seems some of those might<br /> be worth including in the main text. In contrast, the results in Fig 4 are not<br /> quantitative, and overall not that compelling. The top row<br /> of panels are kind of useful in demonstrating you get something like a circle.<br /> but we don't actually know that this corresponds to cell cycle from this picture<br /> (unless I misunderstood, the colors are inferred, not known).<br /> And looking at the mPancreas results one might be tempted to use (-3,0) as the<br /> center of the circle, which would change computation of polar angle quite a bit.<br /> Is there reason to think that sticking with (0,0) is better? If so, any idea why does<br /> the circle show this shift? (Similar issues arise, to a lesser extent, with HippNPC).<br /> The Top2A results are, on their own, too noisy to be convincing -- why not show R2 plots for<br /> all cell-cycle genes (which could be contrasted with non-cell-cycle genes, and also compared<br /> with other methods). And as far as I can<br /> see Fig 4c is, at best, only interesting once one is convinced that the cell cycle<br /> is being correctly inferred -- nothing here to say that the cell cycle inferences are accurate.<br /> To be clear, I'm not saying the method does not generalize well across<br /> data sets; I'm saying that the evidence for this needs to be more clearly presented.

3. A less fundamental issue: I don't really think describing this as an example of "transfer learning"<br /> is helpful. Indeed it is not even clear to me it is accurate.<br /> For example, in the cited Pan et al 2008, they describe the transfer<br /> learning problem as follows: "In a transfer learning setting, some labeled data Dsrc are<br /> available in a source domain, while only unlabeled data Dtar<br /> are available in the target domain." That does not apply here - everything<br /> is based on unlabelled data.

More generally, giving the approach a name like "transfer learning" seems to<br /> suggest that there is something going on to actually make this transfer<br /> from one dataset to another, or some deeper theoretical reason to think it should work<br /> -- but I don't believe either of these is true. You are just hoping<br /> that the PC weights learned in one (carefully chosen) data set will<br /> also work to capture cell cycle on other data sets.<br /> It isn't obvious in advance that this rather simple approach<br /> would work well, and the major contribution of the paper is to assess this<br /> empirically.

1. The abstract is hyperbolic. "ubiquitous applicability of transfer learning";<br /> "can predict any cell's position in the cell cycle",<br /> "universally accurate", "eminently pertinent"...

Minor:

• p2 you introduce the term "cell cycle pseudotime" only to explain later that it is not really a time at all. Why not just go straight into "cell cycle position"<br /> or "cell cycle phase"? (Also, the term "wall time" may not be familiar to all readers?)

-p5 left column: Figure 2d-> 2f?

• p8, right column: is "superficial" the right word here?

• Some of the loess fits (eg Fig 2 d-f; Fig 4 panel b, especially mHippNPC) don't look visually very good. Is this<br /> just an artifact of having 0s, whose density is impossible to see due to overplotting, or is loess over-smoothing? Might trend filtering, as used in Hsiao et al, work better?

Refs:

J Novembre and M Stephens. Interpreting principal component analyses of spatial population genetic variation.<br /> Nat Genet 40(5):646-649, May 2008A Review of Zheng et al, Universal prediction of cell cycle position using transfer learning, by Matthew Stephens

This paper provides a new approach (tricycle) for predicting the<br /> position of a cell in the cell cycle. The approach claims to work<br /> regardless of cell type, species and sequencing assay.<br /> There are several things to like about the paper. In particular,<br /> the tricycle method is very<br /> simple: i) compute the first two PCs on<br /> 500 annotated cell-cycle genes in a data set where cell cycle<br /> is the primary source of variation; ii) project<br /> any future observations to this 2-d embedding and compute<br /> the polar angle to predict its cell cycle<br /> position. Further, the empirical results are promising.<br /> At the same time I think the paper<br /> could be substantially improved by removing or<br /> reducing some of the less innovative parts, toning down some of the rhetoric,<br /> and focussing on the most convincing empirical results. My comments expand<br /> on these suggestions.

Main comments:

1. I found most of the material on PCA not to be<br /> especially novel or interesting. The use of PCA to determine cell cycle<br /> position has a long history (including many papers cited here),<br /> and existing mathematical results already go far beyond<br /> the analysis presented here. The behavior of PCA on cyclic phenomena<br /> is much more general than presented here, and does not rely on sinusoidal<br /> functions or "two distinct peaks" etc. Rather it stems<br /> from the result that cyclic phenomona lead to circulant covariance matrices,<br /> and all circulant matrices have the same eigenvectors:<br /> the columns of the discrete Fourier transform matrix. The result<br /> is that, when the covariance patterns primarily reflect cyclic phenomoena,<br /> the first two PCs will form a circle/ellipse.<br /> See Novembre and Stephens (2008) and references therein for further discussion.<br /> Figure 1 is useful for summarizing the method, but most of the other<br /> material could be condensed or removed and I think the paper would be improved because<br /> it would better focus on what is actually new and interesting, the tricycle<br /> method (currently not introduced until p6) and the empirical assessments of its performance.

2. The paper left me asking myself this: what is the strongest empirical support that tricycle cell<br /> cycle assignments work in practice? To me, Fig 5 panels c and g are the most convincing, because they are quantitative<br /> comparisons with an alternative technology (and one that is often considered the<br /> "gold standard" in this area). I also liked the quantitative comparisons with other<br /> methods, and it seems some of those might<br /> be worth including in the main text. In contrast, the results in Fig 4 are not<br /> quantitative, and overall not that compelling. The top row<br /> of panels are kind of useful in demonstrating you get something like a circle.<br /> but we don't actually know that this corresponds to cell cycle from this picture<br /> (unless I misunderstood, the colors are inferred, not known).<br /> And looking at the mPancreas results one might be tempted to use (-3,0) as the<br /> center of the circle, which would change computation of polar angle quite a bit.<br /> Is there reason to think that sticking with (0,0) is better? If so, any idea why does<br /> the circle show this shift? (Similar issues arise, to a lesser extent, with HippNPC).<br /> The Top2A results are, on their own, too noisy to be convincing -- why not show R2 plots for<br /> all cell-cycle genes (which could be contrasted with non-cell-cycle genes, and also compared<br /> with other methods). And as far as I can<br /> see Fig 4c is, at best, only interesting once one is convinced that the cell cycle<br /> is being correctly inferred -- nothing here to say that the cell cycle inferences are accurate.<br /> To be clear, I'm not saying the method does not generalize well across<br /> data sets; I'm saying that the evidence for this needs to be more clearly presented.

3. A less fundamental issue: I don't really think describing this as an example of "transfer learning"<br /> is helpful. Indeed it is not even clear to me it is accurate.<br /> For example, in the cited Pan et al 2008, they describe the transfer<br /> learning problem as follows: "In a transfer learning setting, some labeled data Dsrc are<br /> available in a source domain, while only unlabeled data Dtar<br /> are available in the target domain." That does not apply here - everything<br /> is based on unlabelled data.

More generally, giving the approach a name like "transfer learning" seems to<br /> suggest that there is something going on to actually make this transfer<br /> from one dataset to another, or some deeper theoretical reason to think it should work<br /> -- but I don't believe either of these is true. You are just hoping<br /> that the PC weights learned in one (carefully chosen) data set will<br /> also work to capture cell cycle on other data sets.<br /> It isn't obvious in advance that this rather simple approach<br /> would work well, and the major contribution of the paper is to assess this<br /> empirically.

1. The abstract is hyperbolic. "ubiquitous applicability of transfer learning";<br /> "can predict any cell's position in the cell cycle",<br /> "universally accurate", "eminently pertinent"...