On 2020-04-02 01:38:27, user Binbin Chen wrote:

This is a great follow-up study from our earlier paper! https://www.biorxiv.org/con...

On 2020-04-02 01:16:57, user Nara wrote:

One needs to take this preliminary results with a grain of salt. The conclusions have for reaching impact. However, the results are based on one strain. Certainly not enough to draw any Major conclusion. The authors would have been well off comparing 4 to 6 strains before rushing to publish even in a preprint server. Extrapolating these results into lower virulance for the Indian strain is misleading. Publication like this can cause a sense of false hope.

On 2020-03-29 15:11:48, user mksharma62 wrote:

Well, this is just one new study and interpretation but on this basis we cannot relax and take chances. Examples of America, Italy and Spain should alert us more.

On 2020-03-28 17:15:09, user Gowrisankar Rajam wrote:

All are very valid questions and observations. Let us be careful in how this non-peer reviewed genomic analysis with very low n (one strain/location?) is interpreted and circulated in the social media. Already there are tweets that interprets this manuscript that the miRNA is mutating the SARS-COV2 in India, making it less virulant to Indians and making them immune which is a gross misinterpretation of this data. We don't want to give false sense of safety to people in India and elsewhere that can lead to defiance of quarantine and social distancing with catastrophic outcomes. Would request you all to share your scientific wisdom in the social media to ensure we handle this pandemic in a scientific and more responsible manner.

On 2020-03-27 13:48:28, user Varun Sharma wrote:

Need a clarification, the novel variation that was observed by authors in the spike protein can be used for the detection.

On 2020-03-27 06:38:20, user Saurabh Mahajan wrote:

This study has many FATAL FLAWS which makes their major conclusion WRONG.

1. As pointed out by many people,THERE IS *NO* INDIAN STRAIN of CoV-2. Tourists and hundreds of returning Indians must have carried different strains from everywhere.

2. Only 2 strains from India have been sequenced till date and submitted to public databases (itself a major lacunae). Strains from 2 patients from Kerala (who arrived from Wuhan) were sequenced in late January and submitted. This study uses ONLY ONE of those strains. Even the second Kerala strain does not share same mutations as the one studied here. Claiming that these are Indian strains is MEANINGLESS.

3. The study also uses only single strains from other countries. It is not clear how these strains were chosen. This also makes the comparisons useless.

4. The phylogenetic methods used here (NJ) are simplistic. Together with low sampling, simplistic methods may have led to obviously incorrect results (Fig 2b). These results can be compared to results from nextstrain.org, which analyses hundreds of genomes in real-time and already includes the Indian genomes.
5. The Indian genome sequences are very different from most other Wuhan strains. This could indicate ERRORS IN SEQUENCING, which are not uncommon.
6. There are several other problems with the computational analysis and its interpretation.

Comparing sequences and studying evolution can give many useful insights. But this can be achieved only with extensive sampling, accurate sequencing methods, and use of rigorous methods. Unfortunately Indian agencies have published/sequenced only 2 strains.

Saurabh Mahajan<br /> Asst Professor<br /> St. Joseph's College, Bangalore

On 2020-03-26 16:03:24, user Vijaykumar Muley wrote:

What is the chance that there is only one viral strain entered into India with a specific mutation, and circulating in population ? Are the number of analyzed genomes enough to support the authenticity of the mutation ?

On 2020-03-25 19:33:23, user Rohit Satyam wrote:

Hi Authors. <br /> Referring to line "We also used psRNATarget server to compare the predicted targets by the<br /> two methods"

I am unable to understand why you have used a plant small RNA target analysis server, psRNATarget in your study?@COVID-19 is not a plant virus.

I hope there are many differences in the prediction as mentioned https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3125753/

"For example, an animal miRNA generally requires loose complementarity in about first eight nucleotides of the miRNA, while a plant miRNA demands the whole miRNA mature sequence to be near perfectly aligned with its mRNA target. Secondly, an animal miRNA tends to inhibit target gene’s expression at the translational level, whereas a plant miRNA directly cleaves its target transcript"

That would make your predictions wrong.

On 2020-03-25 02:26:50, user Yugandhar Reddy wrote:

Can the authors clarify how many SARS-CoV2 Genomes from India were used in the analysis. It is appears in the materials and methods only one genome sequence was analysed. Can you please confirm if only "ONE" sequence from India was analysed?

Thanks<br /> Yugadhar

On 2020-04-02 00:00:54, user Aaron wrote:

The authors should be careful in making a claim that the expression of ACE2 in cats and dogs would allow them to be links in transmission chains. While they discuss sequence similarity, they point out that even the closest ACE2 protein to humans still has 15% divergence. Regardless of receptor protein presence, a virus needs much more than a receptor to replicate within a host. Until there is biological evidence supporting such a claim, this should be removed at the very least from the abstract to avoid misleading future readers.

On 2020-04-01 23:09:16, user Amanda Renkema wrote:

So cats can transmit COVID-19 to eachother, but can they transfer it to humans?

On 2020-04-01 17:43:12, user Lea wrote:

This article is not peer reviewed but now started to be quoted... There are both ethical and methodological questions to this 'study' - and do you realize that putting it in this form, without a peer review, can cost domestic animlals lifes???

On 2020-04-01 22:41:28, user Manuel Kleiner wrote:

Thanks for this very well done and validated study. I like the entrapment database approach that you used. <br /> I personally would emphasize the outcome of your study differently in the discussion though. I think it is actually highlighting the fact that both the traditional two-step approach and the new sectioning approach produce unpredictable FDRs, while the large database search approach gives exactly the expected 1% FDR. Even though the sectioning approach lowers the FDRs in the test dataset, it does not do so in a predictable fashion i.e. we will have no idea what the true FDRs will be in a complex environmental metaproteome.<br /> So I think it would be great if the manuscript could provide a more weighted discussion of the results. If the main goal of a study is to achieve maximal PSM numbers then the sectioning approach is the way to go. If one is looking to get the most reliable and accurate results the traditional large database approach is the way to go. I do think though that for most studies reliably controlled PSMs are needed and that one has to chose to loose out on some extra PSMs to get accurate results.

Of course this all only matters if one analyzes the data on the peptide level, as likely protein inference will change the outcome in terms of FDRs for the different approaches significantly.

On 2020-04-01 19:11:14, user Alex Crits-Christoph wrote:

I thank the author for sharing this work. It is important to maintain a high degree of skepticism and rigor in microbial ecology and the author should be lauded for applying taking the time to apply these principles to a dataset.

However, in my reading of this work, there were a few questions and/or issues arising:

1. on Line 56 - I felt as though the methodology and data used in this line, essential to the interpretation of the manuscript, are a bit too vague to reproduce properly and should be explicitly defined in this manuscript. First, I think that there should be more information given about the database used in this analysis. Looking closely at [ref 3], it appears to be a fairly comprehensive combination of both the SEED and several metagenome assembled genomes from previous works. I'd like to say I strongly appreciate that the author has integrated data from multiple genome-resolved studies, which is essential for characterizing most ecosystems and a huge step above databases like SEED or RefSeq. That being said, I wish more information was provided on this database - can it be hosted somewhere? Can there be a supplementary table listing which genomes are contained in it? I also think this work should separately explain how that database was constructed in the Methods.

On to the methodology - I would assume that the methodology is exactly as described in [ref 3], which is reproduced below:

"Our criteria for assigning homologous ORFs with BLASTp were bit score greater than 50, minimum alignment length of 30 amino acids, minimum sequence similarity of 30%."

I am concerned about whether these methods recapitulate taxonomy accurately. Many (most?) proteins in microbial genomes do not have strong phylogenetic signals (which is why we use marker genes to infer taxonomy) and BLAST "best hits" in the range of 30-70% will not be accurate enough to infer genus level taxonomy, or often higher taxonomic ranks. I would recommend instead performing this taxonomic analysis on just protein sequences known to have strong phylogenetic signal, reasonably netural evolutionary pressures, and assemble well from metagenomes (e.g., ribosomal proteins). The comparison versus all blast hits would be interesting to say the least. I think minimum sequence similarities should be increased to at minimum 60% to infer Phylum/Order taxonomic ranks and possibly 70-90% to infer genus taxonomic ranks as is done in the text, although these numbers are variable (and the 'correct' answer differs for each protein family, of course).

1. General comments on Fig 1:

-- What percentage of reads, and what percentage of assembled transcripts, in each sample were not assigned taxonomy? This is critical for interpretation.

-- Can the author suggest an explanation for why eukaryotic phytoplankton are seen so strongly in the assembly analysis, even in the kit control? What precedent is there for eukaryotic phytoplankton in kit controls? What degree of gene content and coverage is there for these hits? Are there really a large and diverse number of ORFs from eukaryotic phytoplakton in negative controls? What the percentage identities? This seems like an unusual finding.

-- Analyses of contaminating sequences often rely on those sequences being at higher abundances (or higher total counts) in controls, as opposed to biological sequences. (For a wonderful overview and possibly useful software package from the 16S field, see: https://microbiomejournal.b...

However, the author has not done that here, and curiously, the Pseudomonas sequence content substantially increases in the biological samples as opposed to the negative controls. Can the author offer an explanation for this pattern? I really like Figure 1C, which does show that the Pseudomonas data in crustal samples looks similar to that in controls, but I think the author should show more analysis showing that the Pseudomonas detected in the biological samples are basically exactly those detected in the controls, as opposed to true environmental Pseudomonas. This could be done with a BLASTP comparison between Pseudomonas sequences (ideally, taxonomic marker proteins) in the environmental sequences and the controls. For a (probably overly detailed, but very interesting) analysis of a contaminant in genomic data I like to recommend this paper:

https://mbio.asm.org/conten...

Overall, I agree with the author's central claim that contaminating sequences have confounded the biological conclusions of Li et al. 2020. However, I walk away concluding that we have not yet presented adequate genomic clarity on exactly how and why this happened. I believe that further and potentially more careful genome resolved methods can shed some the necessary light on this matter. I commend the author for taking the first steps in this direction.

On 2020-04-01 15:10:57, user Maria Llamazares wrote:

https://uploads.disquscdn.c...

very interesting manuscript!<br /> In a first glance seems that Table 1 labelling is incorrect since you say you are showing the PPE/saline ratio and actually you are doing the other way around: Saline/PPE. Could you double check?

On 2020-04-01 14:58:49, user mov wrote:

Because in all the images the distribution and grouping of infection is in groups of 5 + 1. Represents an order or shield based on viral action?

On 2020-04-01 10:38:30, user Dana MacGregor wrote:

This paper has now been accepted into Plant Physiology in a Letters format. As soon as I have the DOI I will link it here.

On 2020-04-01 08:45:54, user Liz Miller wrote:

This paper was the subject of the Miller lab weekly journal club and, following a fun discussion of the findings, we have the following comments to make. Please bear in mind that we do not study mRNA regulation or decay, but enjoyed reading the manuscript which was somewhat out of our normal area of expertise.

This paper uses a synthetic mRNA library to examine how changes in a 10 nucleotide sequence preceding an upstream open reading frame (uORF) can alter translation and mRNA stability of both the uORF and downstream ORF. The main findings are that translation of the main ORF is protective to the mRNA, whereas improved translation of the uORF is detrimental to the stability of the mRNA. Additionally, sequences are found that prevent translation entirely, whilst others are shown to enable cap-independent translation of the main ORF. Finally, the degradation pathways that these types of messages might follow are identified. Overall, this seems a powerful tool to study the sequence landscape that is possible for modulating uORF and ORF translation and the effects on subsequent mRNA decay. We were excited to think about how the information in this study could be mapped onto the human genome to understand how frequent the different classes of modulatory sequences are and whether any of these have actually been selected for in evolution (one example, that of an RG4 upstream of NRAS is an exciting first step for this!).

We have some short points of note following our group discussion:

1. The text relating to Fig 1b states that there is a reciprocal relationship between tracer peptide (uORF) and GFP (main ORF) abundance (ie. if tracer peptide is translated, GFP is not). This relationship wasn’t apparent to our eyes in the flow cytometry of Fig 1b, which showed that with the AUG condition, increased tracer peptide was generally accompanied by increased GFP. This is probably due to the plasmid-based system used for expression, since in Supplemental Figure 1b. that used RNA transfection, the effect was clear. We spent quite some time puzzling over this (and some of the other flow cytometry plots), so for a naive reader, additional explanation here might be good.

2. In some experiments presented as violin plots (eg. Figure 2B and C), n is clearly very large and the p-value presented is impressive, but we wondered if it would be more useful to plot effect size rather than p-value for these data. With such large datasets, effect size can be more meaningful in understanding how two populations vary. On a similar note, we thought it would be valuable to include the p-values (and effect size) for the in vitro comparisons of mRNA stability to support the claim that the stability is only different between the populations in vivo.

3. Finally, on a stylistic note, we felt that there was a lot of important data in the supplementary figures, in particular the various experiments that follow up on the potential decay mechanisms for these reporters. The short format does not do justice to the full story!

Thank you for sharing your work on BioRXiv and we hope our comments might be useful

On 2020-04-01 06:15:19, user Artem Nedoluzhko wrote:

Thank you for the interesting preprint, In my opinion, several words about sampling (i.g. about Second Fram Expedition) should be added. It will make the manuscript more informative. In another case, Figure 2C looks unclear for people, who don't know about the heroic history of Norwegian Arctic Expeditions. Good luck!

On 2020-04-01 03:31:51, user Tom Terwilliger wrote:

Hi Iris Young, Alexis Rohou and James Fraser,<br /> Thanks for the helpful review of the previous version! We revised the paper and uploaded this version with changes addressing your comments.

Here are some responses to some of your specific comments (already sent them to you but posting here so that anyone can see them).

1. The smoothed squared map procedure is used to get a mask as you note in your comments, but it is not used in the map-phasing procedure beyond that.

2. One of your questions was about further validation. We have recently made a couple new tutorials on density modification. One takes a high-resolution apoferritin dataset and docks a homologous structure into it, fixes loops, and refines. You can run this tutorial with a script that is included. You could also edit this script if you want to use a non-density-modified map (an auto-sharpened one) and carry out the same procedure otherwise. When I do this I find that the density -modified map agrees better with both its own model and with the model from non-density-modified map than the original map does. Further the density-modified map and its model agree best. Here "agrees" means (1) set b values of a model to zero. (2) run map_model_cc on map and model and get CC_mask. This is the best example I have where density modification makes a clear difference in the final model. CC_mask values for this little experiment:<br /> analysis: <br /> denmod_map/denmod model 0.8698 denmod_map/std model 0.8461 std_map/denmod model 0.8114 <br /> std_map/std model 0.8244

3. You can download an excel worksheet that should have the data and figures to answer some of your questions from the bioRxiv web site.

4. If you go to the nice "full text" version of the paper on the bioRxiv site and go to Fig. S6 and click the magnifying glass then you get the plots that you request in the review of things like EMRinger scores. What you will see is that they vary, but are not improved on average by density modification. All the details are in the excel spreadsheet too.

5. "What was the FSC improvement for the 1.8 Å map shown in Fig1C,D?"

This is now clarified:

D. Density-modified map. The estimated improvement in resolution where FSCref is ½ is 0.05 Å.

1. "For the density-modified map, values of estimated Fourier shell correlation to a true map, FSC_ref, are estimated from resolution-dependent error estimates.”. This sentence makes it sound as though FSC_ref is computed not from the FSC between half maps by using eqn 1, but by some other means. Is this so? If so, please expand since FSC_ref seems everywhere else in the main text to be from eqn 1. Perhaps the authors are referring to eqn 9 here, but it is not clear why this is necessary. Why not use FSC between density-modified half maps and eqn 1?."

This is now clarified:

For the density-modified map, values of estimated Fourier shell correlation to a true map, FSCref, are estimated from resolution-dependent error estimates (Eq. 9, Methods). These error estimates are used so that any correlations among the two original half maps and the two map-phasing half maps can be taken into account (see Methods).

1. Typos and clarifications fixed...in particular:

A local weight for each set of half-maps (original half-maps, F1a, F1b and map-phasing half-maps, F2a, F2b) is then obtained as the inverse of the corresponding local variance map. These local weights are then scaled to yield an average local weight of unity and then are applied to the individual half-maps before they are averaged.

1. The reason not to use Eq. 1, which is the standard estimate of map quality:

FSCref =[2 FSC /(1+ FSC)]½

at the end of density modification is that the two half-maps at this point may have some correlation (the value of A in Eq. 5b may not be zero).

If the value of A were zero, then you could use Eq. 1. In practice... it might not make that much difference, but I have not done a systematic investigation.

-Tom T

On 2020-03-18 17:02:11, user Iris Young wrote:

The purpose of this paper was to demonstrate the application of density modification to cryoEM. The authors have previously implemented density modification for macromolecular crystallography, and since then it has been widely used in the field of structural biology. The extension of this method to cryoEM can be expected to be similarly impactful, as several users have already attested to.

The authors validate their approach with apoferritin maps refined to 1.8 and 3.1 Å, based on the premise that the 3.1 Å map should better resemble the 1.8 Å map following density modification. They also use this test case to validate a metric for map resolution inferred from the agreement between two half-maps, enabling them to assess resolution after each iteration of density modification; a small change in resolution between cycles serves as the convergence criterion. We would be interested to see further validation of the utility of density modification in the form of improvement of models built into a map as judged by RMS-Z of bond lengths or angles, EMRinger score (https://www.nature.com/arti..., Q-scores (https://www.biorxiv.org/con..., or other metrics used by molprobity (https://www.ncbi.nlm.nih.go... or model validation by the wwPDB (https://www.ncbi.nlm.nih.go....

This paper is successful in laying the necessary groundwork for the concept of density modification, outlining what concepts can be carried over directly from crystallography and what assumptions are broken in the case of electron microscopy, and detailing the authors' implementation of density modification for cryoEM in the Phenix software suite. Combined with the references to other foundational theoretical work, the paper should provide enough information to understand the procedure in sufficient detail to reimplement it.

We found it took several read-throughs and group discussions before differences in understanding could be resolved, however, and feel the paper would benefit from a flow chart similar to our figure below. In brief, our understanding of the algorithm is as follows:

https://uploads.disquscdn.c...

1. Half maps are trimmed to a volume containing the particle and not too much extra.
2. A starting map is derived from this pair of half maps by averaging, then blurring or sharpening to a resolution calculated from the agreement of the half maps.
3. The starting map is thresholded to whatever value produces a mask encompassing a total volume matching the expected particle volume.
4. Intensities in the starting map are used to generate histograms for the masked (protein) and unmasked (solvent) regions.
5. Another pair of smoothed squared maps are generated from the starting half maps, representing the local variation in electrostatic potential.
6. These maps and the histograms are used to predict the most likely value of each structure factor in reciprocal space half maps, producing "map-phasing" maps.
7. Density modified maps are generated by taking a weighted average of the reciprocal space starting half maps and map-phasing half maps. Weights are chosen in shells of resolution based on the correlation between the starting half maps.
8. A new resolution is estimated based on the agreement of the density modified maps. If the resolution has improved substantially, the density modified maps are reused as starting half maps in another iteration. In the current implementation, histograms are not regenerated every cycle, only map-phasing maps.
9. Upon convergence, the density modified half maps can be averaged to produce a new full map.

The few limitations of this manuscript are matters of clarity and readability. For example, in our first group discussions, readers did not immediately grasp that the map-phasing map was calculated in reciprocal space. Several important concepts and equations are buried two or three references deep, at which point there is ambiguity about how directly they translate to the cryoEM case. Reiterating a few more details in the current manuscript would aid understanding here and would increase the long-term usefulness of the manuscript. For example, the description of the squared smoothed map Z(x) and equation 3 in reference 5 (or a sum-up of this in words) would be helpful.

Perhaps most importantly, it was not clear to us that F2, referred to at several points as a density-modified map, is the same as the map-phasing map and that G is the first recombined density-modified map. This could be clarified either in words or by a different choice of variable names. The description of a weighted averaging step could directly reference equation 7a. It would also be helpful to clearly state which steps take place in real space vs. reciprocal space and by individual structure factors or in shells of resolution, or to indicate these in a diagram or flow chart.

Some sections could bear to be expanded and broken into smaller pieces. For example, the sentence beginning "The overall procedure for density modification of two half-maps" is quite long and quite dense, and step (4) in particular is difficult to parse. Around the descriptions of some calculations, it is at times difficult to follow either the direction of derivation of terms or the direction of flow of execution in the current implementation; separating these more explicitly would be helpful.

Overall we are very enthusiastic about this work. The extension of density modification to cryoEM comes at a critical time for the field, when the number of structures determined by cryoEM is increasing exponentially and resolutions now frequently permit atomistic interpretation. This manuscript clearly articulates the differences between the crystallography and EM cases and how these are taken into consideration. Additionally it describes several possible alternative choices for other implementations, such as using a mask that distinguishes between more than two regions or how to do density modification when half-maps are not available. We thank the authors for posting a preprint so that the algorithm can be immediately understood and reimplemented.

Minor points:<br /> • What was the FSC improvement for the 1.8 Å map shown in Fig1C,D?<br /> • Fig 2B: The scatter plot is useful, but we would like to see a histogram of the delta in average map-model FSC, because this would tell us, for example, how frequently the average FSC can actually get worse and by how much. This could be done, e.g. for maps in the 2.8-3.3 range, or some other chosen range within which there is a sufficient number of samples. Such histograms would be useful to early users of the program, so that they know what to expect in a general sense.<br /> • Similar could be done for Fig 2A.<br /> • “For the density-modified map, values of estimated Fourier shell correlation to a true map, FSC_ref, are estimated from resolution-dependent error estimates.”. This sentence makes it sound as though FSC_ref is computed not from the FSC between half maps by using eqn 1, but by some other means. Is this so? If so, please expand since FSC_ref seems everywhere else in the main text to be from eqn 1. Perhaps the authors are referring to eqn 9 here, but it is not clear why this is necessary. Why not use FSC between density-modified half maps and eqn 1?. <br /> • “Another is that half-maps with relatively independent errors are available in electron cryo microscopy but not crystallography”: “relatively independent” is a confusing phrase. Perhaps the authors mean something like the “nominally independent”, or “independent in principle” or the like?<br /> • There is a typo in the line "where <βa2> = <βa2> = B" -- the second "a" should be "b."<br /> • Equation 6d should define S, not D.<br /> • There appears to be one too many instances of the word "averaged" in the sentence "Where w is the weight on the original half maps (F1) and (1-w) is the weight on averaged density modified averaged half-maps (F2) and the averaged maps are given by:"<br /> • Equation 9 would be much more readable as: FSCref = 2(2A + B + S)(2A + B)(C + S + 2) + (C + 2)S<br /> • Several equations would be much more readable if formatted as fractions instead of with "/" and layers of parentheses.<br /> • In the real-space weighting section, it would be helpful to explicitly label which half-maps, original or modified, are being handled in each step.<br /> • Most instances of subscripts on FSC have an unnecessary leading underscore (e.g. FSC_d_min, as do all instances of d_dm.

We review non-anonymously and have posted this comment on the preprint at BioRxiv, Iris Young, Alexis Rohou and James Fraser.

On 2020-04-01 00:09:30, user Sinai Immunol Review Project wrote:

The authors of this study sought to characterize the immune mechanism causing severe pulmonary disease and mortality in 2019-nCoV (COVID-19) patients. Peripheral blood was collected from hospitalized ICU (n=12) and non-ICU (n=21) patients with confirmed 2019-nCoV and from healthy controls (n=10) in The First Affiliated Hospital of University of Science and Technology China (Hefei, Anhui). Immune analysis was conducted by flow cytometry. 2019-nCoV patients had decreased lymphocyte, monocyte, and CD4 T cell counts compared to healthy controls. ICU patients had fewer lymphocytes than non-ICU patients. CD4 T cells of 2019-nCoV patients expressed higher levels of activation markers (OX40, CD69, CD38, CD44) and exhaustion markers (PD-1 and Tim3) than those of healthy controls. CD4 cells of ICU patients expressed significantly higher levels of OX40, PD-1, and Tim3 than those of non-ICU patients. 2019-nCoV patients had higher percentages of CD4 T cells co-expressing GM-CSF and IL-6 compared to healthy controls, while ICU patients had a markedly higher percentage of GM-CSF+ IFN-γ+ CD4 T cells than non-ICU patients. The CD4 T cells of nCoV patients and healthy controls showed no differences in TNF-α secretion.

The CD8 T cells of 2019-nCoV patients also showed higher expression of activation markers CD69, CD38, and CD44, as well as exhaustion markers PD-1 and Tim3, compared to healthy controls. CD8 T cells of ICU patients expressed higher levels of GM-CSF than those of non-ICU patients and healthy controls. No IL-6 or TNF-α was found in the CD8 T cells of any group. There were no differences in numbers of NK cells or B cells in 2019-nCoV patients and healthy controls, nor was there any GM-CSF or IL-6 secretion from these cells in either group.

Percentages of CD14+ CD16+ GM-CSF+ and CD14+ CD16+ IL-6+ inflammatory monocytes were significantly increased in nCoV patients compared to healthy controls; in particular, patients in the ICU had greater percentages of CD14+ CD16+ IL-6+ monocytes than non-ICU patients. The authors suggest that in 2019-nCoV patients, pathogenic Th1 cells produce GM-CSF, recruiting CD14+ CD16+ inflammatory monocytes that secrete high levels of IL-6. These may enter pulmonary circulation and damage lung tissue while initiating the cytokine storm that causes mortality in severe cases. This is consistent with the cytokine storm seen in similar coronaviruses, as IL-6, IFN-γ, and GM-CSF are key inflammatory mediators seen in patients with SARS-CoV-1 and MERS-CoV.

Limitations of the study: Though the results of this study open questions for further investigation, this is an early study on a small cohort of patients, and as such there are a number of limitations. The study included only 12 ICU patients and 21 non-ICU patients, and ideally would be repeated with a much larger patient cohort. Though the authors make claims about differences in lymphocyte and monocyte counts between patients and healthy controls, they did not report baseline laboratory findings for the control group. Additionally, severity of disease was classified based on whether or not patients were in the ICU. It would be interesting to contextualize the authors’ immunological findings with more specific metrics of disease severity or time course. Noting mortality, time from disease onset, pre-existing conditions, or severity of lung pathology in post-mortem tissue samples would paint a fuller picture of how to assess risk level and the relationship between severity of disease and immunopathology. Another limitation is the selection of cytokines and immune markers for analysis, as the selection criteria were based on the cell subsets and cytokine storm typically seen in SARS-CoV-1 and MERS-CoV patients. Unbiased cytokine screens and immune profiling may reveal novel therapeutic targets that were not included in this study.

Significance: This study identifies potential therapeutic targets that could prevent acute respiratory disease syndrome (ARDS) and mortality in patients most severely affected by COVID-19. The authors propose testing monoclonal antibodies against IL6-R or GM-CSF to block recruitment of inflammatory monocytes and the subsequent cytokine storm in these patients.

On 2020-03-31 23:13:00, user Aaron wrote:

I have concerns with the authors' claim that the US has the most evolved viral population. The study only includes 95 genomes, 54 of which are from the US. Of the remaining countries, six of them (Australia, Brazil, Italy, Nepal, South Korea, and Sweden) are represented by only a single genome, two (India and Taiwan) are represented by two genomes, Japan is represented by three genomes, and China is represented by 30 genomes. These sample sizes are vastly too small to be making any claims of viral diversity or patterns of spread. While the NCBI database has been rather limited in the number of available genomes (only 320 full genomes available as of March 31, 2020), there are many, many more available through GISAID's Next HCoV19 App. If this resource is an option for the authors, I'd strongly recommend redoing the phylogeographic analyses to include the sequences available from GISAID, otherwise there is insufficient power to be making these claims.

On 2020-03-31 22:43:44, user Alexander Ljubimov wrote:

The revised version of this paper has been published in Nature Communications: Anna Galstyan, Janet L. Markman, Ekaterina S. Shatalova, Antonella Chiechi, Alan J. Korman, Rameshwar Patil, Dmytro Klymyshyn, Warren G. Tourtellotte, Liron L. Israel, Oliver Braubach, Vladimir A. Ljubimov, Leila A. Mashouf, Arshia Ramesh, Zachary B. Grodzinski, Manuel L. Penichet, Keith L. Black, Eggehard Holler, Tao Sun, Hui Ding, Alexander V. Ljubimov & Julia Y. Ljubimova. Blood–brain barrier permeable nano immunoconjugates induce local immune responses for glioma therapy. Nature Communications volume 10, Article number: 3850 (2019).

On 2020-03-31 22:15:39, user Mike Rayko wrote:

Can you please double check? In our study (coming soon) we observe deletion in the samples from independent labs at 1605-1607, changing ND (AATGAC) to N (AAC).<br /> Also, Asp268 is 1604-1606 (at least in NC_045512.2)

On 2020-03-26 18:08:36, user overmind wrote:

In the reference genome https://www.ncbi.nlm.nih.go... coordinates 1607-1609 map to GAC (D/Asp), but if you look closely at the bar in Figure1, the area highlighted as 1607-1609 is TGA (off by one)?

On 2020-03-31 22:08:38, user Sinai Immunol Review Project wrote:

Summary and key findings: The authors reported a human monoclonal antibody that neutralizes SARS-CoV-2 and SARS-Cov which belong to same family of corona viruses. For identifying mAbs, supernatants of a collection of 51 hybridomas raised against the spike protein of SARS-CoV (SARS-S) were screened by ELISA for cross-reactivity against the spike protein of SARS-CoN2 (SARS2-S). Hybridomas were derived from immunized transgenic H2L2 mice (chimeric for fully human VH-VL and rat constant region). Four SARS-S hybridomas displayed cross-reactivity with SARS2-S, one of which (47D11) exhibited cross-neutralizing activity for SARS-S and SARS2-S pseudotyped VSV infection. A recombinant, fully human IgG1 isotype antibody was generated and used for further characterization.

The humanized 47D11 antibody inhibited infection of VeroE6 cells with SARS-CoV and SARS-CoV-2 with IC50 values of 0.19 and 0.57 μg/ml respectively. 47D11 mAb bound a conserved epitope on the spike receptor binding domain (RBD) explaining its ability to cross-neutralize SARS-CoV and SARS-CoV-2. 47D11 was shown to target the S1B RBD of SARS-S and SARS2-S with similar affinities. Interestingly, binding of 47D11 to SARS-S1B and SARS2-S1B did not interfere with S1B binding to ACE2 receptor-expressing cells assayed by flow cytometry.

Limitations: These results show that the human 47D11 antibody neutralizes SARS-CoV and SARS-Cov2 infectivity via an as yet unknown mechanism that is different from receptor binding interference. Alternative mechanisms were proposed but these as yet remain to be tested in the context of SARS-CoV2. From a therapeutic standpoint and in the absence of in vivo data, it is unclear whether the 47D11 ab can alter the course of infection in an infected host through virus clearance or protect an uninfected host that is exposed to the virus. There is a precedent for the latter possibility as it relates to SARS-CoV that was cited by the authors and could turn out to be true for SARS-CoV2.

Relevance: This study enabled the identification of novel neutralizing antibody against COV-that could potentially be used as first line of treatment in the near future to reduce the viral load and adverse effects in infected patients. In addition, neutralizing antibodies such as 47D11 represent promising reagents for developing antigen-antibody-based detection test kits and assays.

On 2020-03-26 18:28:55, user Ezi213 wrote:

Try radioactively labeling the virus to identify it, then use that to extract the virus.

On 2020-03-26 01:32:41, user Allison Armstrong wrote:

Since the site that is needed for the virus to be neutralized has been determined, couldn't an aptamer be made to "fit" this site? It could be used for testing, or potentially for a treatment...

On 2020-03-15 01:15:24, user vs wrote:

if everyhting goes well,how long it will take to reach out to patients?

On 2020-03-31 21:48:04, user Brett Pike wrote:

Is CBCAS more likely to produce THC than CBDAS?

On 2020-03-31 21:29:20, user Einar Eftestøl wrote:

Congratulations on an interesting paper. As I have been working in the field of load/mechanotransduction and its effect on skeletal muscle hypertrophy for quite a few years, I would like to recommend reading my paper on the subject from 2016 (PMID: 27488660). Best wishes, Einar Eftestøl

On 2020-03-31 21:14:14, user Jianhua Xing wrote:

The work has been published in Biology Methods and Protocols<br /> https://academic.oup.com/bi...

On 2020-03-31 17:48:05, user Aurel Wünsch wrote:

Useful Proof of Concept! Have you also considered Phenol/Chloroform RNA-Extraction as possible alternative when Kits are not available instead of skipping that part altogether?

On 2020-03-31 15:43:46, user David A.Carlson wrote:

This work was done without using gas chromatography (GC) to find which actual sugars were present. I don't understand why not. So much more detail is possible using GC.

On 2020-03-31 15:19:29, user Azmeraw T. Amare wrote:

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

On 2020-03-31 14:36:50, user NYUPeerReview wrote:

NOTE: This paper was selected for discussion and critique in “Peer Review in the Life Sciences”, a course for PhD students at the New York University School of Medicine. This course aims to build skills in the critical reading of the scientific literature, and provide formal training in the process of peer review. Following discussion as a class, three students wrote this peer review under the guidance of course instructors Damian Ekiert and Gira Bhabha.

The BAM complex is essential for maintenance of the outer membrane of Gram-negative bacteria. It is known that the BAM complex is responsible for export of the outer membrane lipoprotein RcsF. How the BAM complex interacts with its lipoprotein binding partners, like RcsF, at the molecular level is poorly understood. In this paper, the authors determined the crystal structure of the inward-open conformation of BamA bound to RcsF at 3.8 Å resolution. The authors found RcsF lodged in the β-barrel of BamA and identified specific regions of interaction. Using structural comparisons of their model against previously solved conformational states of BamA, the authors found that RcsF binding is incompatible with the outward-open conformation of BamA. Lastly, the authors used crosslinking experiments to demonstrate that cellular levels of BamCDE modulate the formation of BamA-RcsF complexes and may promote the maturation of RcsF-Omp complexes. <br /> Our class was excited by the results of this paper, and had a nice discussion about the implications of this work for the field. Though many details of the structure such as side chains are not clearly defined at 3.8 Å resolution with significant diffraction anisotropy, the model appears to be well-refined and provides useful insights into this important structure. Moreover, the structural interactions between BamA and RcsF were validated using multiple orthogonal methods. The experiments were thorough and appropriate, providing insight into a previously undetermined transport mechanism. Below are some comments that came up during our class discussion, and we hope will be helpful to the authors:

1) While an interesting observation was made with the BamAΔloop1 mutants, the rationale for selecting this loop for deletion was unclear to us. Were other loops also tested but caused loss of Bam activity? It would be nice to know if there is no disruption of RcsF interaction with BamA upon deletion of a different non-essential loop.

2) The authors suggest a model in which the flux of incoming OMP substrates triggers conformational changes in BamA and the release of RcsF to its OMP partners. However, we didn’t see any experiments presented that directly addressed this (e.g., changing Omp expression levels and assessing RcsF maturation in response). Based on the presented data in the paper (notably, where OmpA-RcsF cross-linked product was observed with BamABCDE overexpression, but not BamA or BamAB), we think a major key finding of this work is that RcsF binding to BamA is dependent on BamA conformation and that the BamA conformation may be influenced by the presence of Bam accessory proteins. The triggers of BamA conformational cycling and exchange of RcsF to OMP partners remained unclear to us.

3) In Extended Data Figure 4, we noticed that all the residues chosen for incorporation of the lysine analog are on the inside of the BamA β-barrel. We discussed that including a crosslinker location away from the RcsF binding site that is not expected to crosslink (as a negative control) would strengthen the data and clearly demonstrate the specificity of the assay.

4) Our class had a brief discussion about how the rise of pre-prints may change the general practice in the field of releasing structural data to coincide with publication. As pre-prints increasingly gain visibility and can serve as a means of establishing priority of discovery, it seems worthwhile for the structural biology community to discuss/reassess when structural models and maps should be shared, and perhaps redefine a standard in the field. Should they still be released upon publication in a journal? Or should we be thinking about releasing this data along with the pre-print? Having the coordinates available while reading a structure manuscript can make it much easier to grasp the key points.

5) We found some of the figures a bit difficult to follow. For example, in Figure 3A, it is hard to see the steric clash between the outward-open conformation of BamA and RcsF. Perhaps a different color choice, rendering, and/or orientation of the outward-open state would be helpful for the reader.

6) We noticed that the structure statistic table did not include Ramachandran statistics or CC1/2. We think it would be good to include both of these.

7) We were confused about the nature of the experiment presented in Extended Figures 1B and 1C. Was native gel electrophoresis performed on purified complexes, then the resulting 8 bands excised, and then somehow separated by SDS-PAGE? It was unclear from the main text, legend, and methods sections.

8) The sensogram in Extended Figure 5C is missing labels and not described in legend (i.e. what do the colors represent?).

On 2020-03-31 11:07:28, user James Kirchner wrote:

Interesting study, and a valiant effort at a hard problem.

it is worth noting, though, that the conclusion that peer review doesn't add much to the quality of published papers is vulnerable to two well-known biases (selection bias and survivor bias) which are barely mentioned.

The study can only consider papers that were posted as pre-prints, which might be better than the typical submissions that go to journals (selection bias). In other words, journals may receive lots of weak papers that are rejected during peer review, but unless those weak papers are also posted to pre-print servers, they will never be evaluated in this study. Anecdotal discussions I've had with editors suggest that a large fraction of journal submissions are "dead on arrival", implying that the editorial/peer review process is, in fact, greatly improving journal quality by keeping some of the dreck out.

In addition, the paired-sample comparisons can only consider papers that were posted as pre-prints, and were ultimately published somewhere (survivor bias). Those that were culled out by the peer-review process (or that were never submitted at all) will not appear in these comparisons.

These biases are unavoidable, and while they don't invalidate a study of this kind, they do limit the inferences that should be drawn. In particular, the inference that peer review doesn't improve the quality of science is inconsistent with our everyday experience that merely the expectation that our work will be subject to peer review forces us to be more careful than we might have been otherwise. How many of us have had to tell a grad student, "look, I know you *believe* that, but you the data don't *prove* it so it won't get through peer review"?

The checklist approach used in this study has obvious limitations as well, which the authors briefly discuss but have been largely lost in the subsequent media commentary. This study focuses on whether certain formal items have been documented, but not on the essential question of whether a paper makes sense: do the data justify the conclusions? I have reviewed dozens of papers that suffered from gross methodological, logical, and even mathematical errors. These were detected and corrected through the peer-review process, but none of these issues would appear in the checklists used in this study.

Should preprints "be considered valid scientific contributions"? They are contributions (and potentially valuable ones), but their validity has been vetted only by the authors themselves, who have obvious conflicts of interest. By contrast, peer-reviewed papers have at least been vetted by (in the best case) independent editors and reviewers as well.

Neither process is perfect. But there is a reason by bioRxiv now displays a banner saying<br /> ______________________ <br /> "BioRxiv is receiving many new papers on coronavirus SARS-CoV-2. A reminder: these are preliminary reports that have not been peer-reviewed. They should not be regarded as conclusive, guide clinical practice/health-related behavior, or be reported in news media as established information." <br /> ______________________ <br /> There are crucial differences between the peer-reviewed literature and the blogosphere, whether or not those differences can be detected by studies like this one.

On 2020-03-31 08:03:03, user Liz Miller wrote:

This paper was the subject of the Miller lab weekly journal club and, following a fun discussion of the findings, we have the following comments to make. Please bear in mind that we do not study membrane flippases but enjoyed reading the manuscript which was somewhat out of our normal area of expertise.

The paper combines structural and biochemical analysis of the plasma membrane phospholipid flippase ATP11C in complex with CDC50A, resulting in a final model whereby substrate lipids are first extracted from the membrane and bound in a ‘surface cavity’ before progressing through a lipid exposed crevasse down to a central occlusion site. At this site, the lipid head group is coordinated by conserved residues and prevented from crossing the bilayer by a conserved gate. Collectively, the data develop our mechanistic understanding of P-type membrane flippases.

We have some brief comments that arose following our group discussion:

1. The structural and functional experiments connecting the helix-kinking of P94 to the transport of larger substrates than other P-type ATPases, which typically transport smaller substrates was a very pleasing biological explanation for an intriguing problem.

2. Although other structures of similar transporters in a variety of states including the E2P state observed in this paper are also available, we note the extra insight provided by the fortuitous observation of substrates bound in both the channel and crevasse. This paper serves as a great example of how structures can provide new insight when similar structures of the same conformational state are already available.

3. We note the interesting effect of requiring multiple mutations to reveal a phenotype in the charged TM3/4 loop. In some ways this parallels the requirement to make multiple mutations in the hydrophobic binding groove of various membrane protein insertion machinery (e.g. the GET targeting machinery). Perhaps this may reflect general properties of non-specific interaction surfaces considering that multiple mutations are required to disrupt relatively non-specific binding surfaces of both polar and hydrophobic binding sites.

4. One area we found slightly confusing was the relative orientation of the PS bound to the membrane surface cavity. We were surprised to see the head group of the phospholipid facing away from the crevasse with the acyl chains facing down towards the second binding site. In this arrangement it seems as though the PS would have to slide backwards out of the ‘cavity’ to flip so that the PS head group could fit into the central active site. Obviously we cannot inspect the models ourselves as they are not available on the PDB yet, but we are curious about how the authors envisage this movement occurring without the lipid re-entering the bilayer tail first.

Thank you for sharing your work on BioRXiv and we hope our comments are of some use/interest to you and the wider community :)

On 2020-03-31 02:52:06, user Sinai Immunol Review Project wrote:

Potent binding of 2019 novel coronavirus spike protein by a SARS coronavirus-specific human monoclonal antibody

Keywords<br /> Monoclonal antibody; Cross-reactivity; receptor binding domain

Summary<br /> Considering the relatively high identity of the receptor binding domain (RBD) of the spike proteins from 2019-nCoV and SARS-CoV (73%), this study aims to assess the cross-reactivity of several anti-SARS-CoV monoclonal antibodies with 2019-nCoV. The results showed that the SARS-CoV-specific antibody CR3022 can potently bind 2019-nCoV RBD.

Major findings<br /> The structure of the 2019-nCoV spike RBD and its conformation in complex with the receptor angiotensin-converting enzyme (ACE2) was modeled in silico and compared with the SARS-CoV RBD structure. The models predicted very similar RBD-ACE2 interactions for both viruses. The binding capacity of representative SARS-CoV-RBD specific monoclonal antibodies (m396, CR3014, and CR3022) to recombinant 2019-nCoV RBD was then investigated by ELISA and their binding kinetics studied using biolayer interferometry. The analysis showed that only CR3022 was able to bind 2019-nCoV RBD with high affinity (KD of 6.3 nM), however it did not interfere with ACE2 binding. Antibodies m396 and CR3014, which target the ACE2 binding site of SARS-CoV failed to bind 2019-nCoV spike protein.

Limitations<br /> The 2019-nCoV RBD largely differ from the SARS-CoV at the C-terminus residues, which drastically impact the cross-reactivity of antibodies described for other B beta-coronaviruses, including SARS-CoV. This study claims that CR3022 antibody could be a potential candidate for therapy. However, none of the antibodies assayed in this work showed cross-reactivity with the ACE2 binding site of 2019-nCoV, essential for the replication of this virus. Furthermore, neutralization assays with 2019-nCoV virus or pseudovirus were not performed. Although the use of neutralizing antibodies is an interesting approach, these results suggest that it is critical the development of novel monoclonal antibodies able to specifically bind 2019-nCoV spike protein.

Reviewed by D.L.O as part of a project by students, postdocs and faculty at the Immunology Institute of the Icahn school of medicine, Mount Sinai.

On 2020-03-31 02:51:30, user Alex Crits-Christoph wrote:

I thank the authors for submitting this work. I have a few questions and/or issues arising:

1. The benchmarking performed I think was similar to that of the Almeida 2018 paper [ref 37], but it feels very unclear from the way the current text is written and I think should be specified further - ideally with specific QIIME2 commands that were run.

2. If the benchmarking performed was similar to that of [ref 37], I'd like to give a few different perspectives on potential issues with that benchmarking scheme. That QIIME pipeline is described as:

"Lastly, for QIIME 2 we first dereplicated the query sequences using the vsearch dereplicate-sequences function and then assigned them to the Greengenes (13_8) or SILVA 128 (99% identity clusters) databases using the feature-classifier classify-sklearn function"

This pipeline is confusing to me, because as far as I know, this is not standard practice in most 16S analysis. For examples, it is possible to look at the set of tutorials provided by the QIIME2 developers:<br /> https://docs.qiime2.org/202...<br /> https://docs.qiime2.org/202...<br /> https://docs.qiime2.org/202...<br /> https://docs.qiime2.org/202...

Standard practice is to call OTUs or ASVs ("amplicon sequence variants") across samples (as is done in the tutorials) and then to assign taxonomy to representative sequences. This greatly reduces the number of sequences to classify and restricts classification to high quality sequences. But most importantly, ASVs or de novo OTU clustering allow you to track phylogenetically coherent, statistically independent microbial units across samples that do not rely on reference databases. These units are not reliant on taxonomic classifiers or reference databases - taxonomy is metadata assigned to OTUs, useful for interpretation of results from statistical analysis.

For more on this, see:<br /> https://www.nature.com/arti...

Studying any natural microbial system, it is critically important to be able to independently track microbial species that are not present in reference databases. Therefore, I think it would be most fruitful to do a comparison of kraken and QIIME's classifier on ASV/OTU representative sequences as opposed to raw reads, and compare specificity and sensitivity at various taxonomic ranks.

1. I have another concern with the benchmarking dataset created by Almeida 2018 with regards to taxonomy- specifically, I believe it is mostly (entirely?) composed of genomes from bacterial isolates, which represent a small fraction of bacterial diversity - about 63% of bacterial phyla are *only represented by a genome assembled from a metagenome*: https://gtdb.ecogenomic.org...

In light of this, I am particularly interested in seeing how environmentally common but uncultivated taxa perform for both classifiers specifically, considering that these taxa comprise likely the most serious obstacle for reference-based metagenomic classifiers. A list of some worth checking in particular:<br /> -- TM7 (Saccharibacteria)<br /> -- Candidatus Rokubacteria<br /> -- Candidatus Melainabacteria<br /> -- Candidatus Dormibactereota (AD3)

Thanks to the authors for considering these comments.

On 2020-03-30 23:15:31, user ppgardne wrote:

A few comments:<br /> * how do you compute MCC without a reference structure? It's not clear how you have defined TP, TN, FP & FN for this.<br /> * In addition to the RNAz screen, validating predictions with R-scape (http://eddylab.org/R-scape/):veXHpKUvUAwcwQMpZjJah7PXWxY "http://eddylab.org/R-scape/)") would increase our confidence that these are truly conserved structures. It's a conservative approach, but this seems important here. In a similar vein, alifoldz may be useful too -- RNAz can return a lot of false positives if not used carefully. <br /> *The unstructured regions could also be verified using an energy model such as RNALfold (part of the Vienna package). We have found this a very useful approach for identifying accessible regions in mRNA (https://www.biorxiv.org/con.... <br /> *I was pleasantly surprised to see that some Rfam models matched the sequence. How strong were these matches? Did you use GA thresholds, or drop these in favour of evalues? In fact, how were the comparisons to Rfam models performed? <br /> *Is the methods section missing?

On 2020-03-30 22:42:03, user Thomas Pucadyil wrote:

Very insightful...really enjoyed reading this paper. Congratulations, authors!

On 2020-03-30 21:40:32, user Jessica wrote:

It would be useful if this application was compared to the uORF plugin available for VEP: https://www.biorxiv.org/con...

On 2020-03-30 18:24:04, user Mohamed Diwan wrote:

The title of this paper is misleading and incorrect. 'high-throughput screening' in the title should be replaced with 'virtual screening' or 'docking'. There is no experimental work in this paper.

On 2020-03-30 18:05:15, user Max Crispin wrote:

Thank you Andrew. Spotting omissions like that really demonstrates the value of pre-print servers! We have deposited raw mass spec files in the MassIVE server and are frantically preparing a dossier of all the other files for release. Ahead of that, do please email requests. <br /> Thanks again.

On 2020-03-29 18:25:52, user Andrew Crowley wrote:

Very informative! Will you be depositing the model that you created for figure 3 anywhere? Figure 1 appears to be missing a branched marker at N801.

On 2020-03-30 16:00:00, user Sinai Immunol Review Project wrote:

Summary and key findings: A panel of ~3,000 FDA- and IND-approved antiviral drugs were previously screened for inhibitory efficacy against SARS CoV, a coronavirus related to the novel coronavirus SARS CoV-2 (79.5%) homology. 35 of these drugs along with another 15 (suggested by infectious disease specialists) were tested in vitro for their ability to inhibit SARS CoV-2 infectivity of Vero cells while preserving cell viability. The infected cells were scored by immunofluorescence analysis using an antibody against the N protein of SARS CoV-2. Chloroquine, lopinavir and remdesivir were used as reference drugs.

Twenty four out of 50 drugs exhibited antiviral activity with IC50 values ranging from 0.1-10mM. Among these, two stood ou: 1) the-anti helminthic drug niclosamide which exhibited potent antiviral activity against SARS CoV-2 (IC50=0.28 mM). The broad-spectrum antiviral effect of niclosamide against SARS and MERS-CoV have been previously documented and recent evidence suggests that in may inhibit autophagy and reduce MERS C0V replication. 2) Ciclesonide, a corticosteroid used to treat asthma and allergic rhinitis, also exhibited antiviral efficacy but with a lower IC50 (4.33mM) compared to niclosamide. The antiviral effects of ciclesonide were directed against NSP15, a viral riboendonuclease which is the molecular target of this drug.

Limitations: The drugs were tested against SARS CoV-2 infectivity in vitro only, therefore preclinical studies in animals and clinical trials in patients will be need for validation of these drugs as therapeutic agents for COVID-19. In addition, niclosamide exhibits low adsorption pharmatokinetically which could be alleviated with further development of drug formulation to increase effective delivery of this drug to target tissues. Nonetheless, niclosamide and ciclesonide represent promising therapeutic agents against SARS CoV-2 given that other compounds tested in the same study including favipiravir (currently used in clinical trials) and atazanavir (predicted as the most potent antiviral drug by AI-inference modeling) did not exhibit antiviral activity in the current study.

On 2020-03-30 13:59:05, user Sinai Immunol Review Project wrote:

Summary: Using both publicly available scRNA-seq dataset of liver samples from colorectal patients and scRNA-sequencing of four liver samples from healthy volunteers, the authors show that ACE2 is significantly enriched in the majority of cholangiocytes (59.7 %) but not in hepatocytes (2.6%).

Main findings and limitations : Using bioinformatics approaches of RNASeq analysis, this study reveals that ACE2 dominates in cholangiocytes and is present at very low levels in hepatocytes. The study does not provide mechanistic insights into how SARS-CoV-2 can infect and replicate in cholangiocytes and the types of intrinsic anti-viral responses induced by cholangiocytes when infected. In addition, because the study relies on the assumption that SARS-CoV-2 infects cells only through ACE2, it cannot discount the possibility that the virus can infect hepatocytes through mechanisms other than ACE2-mediated entry. Furthermore, because the scRNA-seq analysis were performed on healthy liver samples, one cannot draw any definitive conclusions about gene expression states (including ACE2 expression in liver cell types) in system-wide inflammatory contexts.

Significance of the finding: This article with other studies on liver damage in COVID patients suggests that liver damage observed in COVID patients is more due to inflammatory cytokines than direct infection of the liver. Even if cholangiocytes are infectable by SARS-CoV-2 (which was demonstrated by human liver ductal organoid study (10.1101/2020.03.16.990317)), published clinical data show no significant increase in bile duct injury related indexes (i.e. alkaline phosphatase, gamma-glutamyl transpeptidase and total bilirubin). In sum, it underscores the importance of future studies characterizing cellular responses of extra-pulmonary organs in the context of COVID or at least in viral lung infections..

Review by Chang Moon as part of a project by students, postdocs and faculty at the<br /> Immunology Institute of the Icahn school of medicine, Mount Sinai.

On 2020-03-30 13:50:05, user Laura Uelze wrote:

Dear Shaoting Li, Shaokang Zhang and Xiangyu Deng,

Thank you for your thorough re-analysis of our data!<br /> We agree that the GC-bias has a major impact on serotyping results and should<br /> be taken into account when analyzing WGS data. We have addressed this issue in<br /> our publication and we will continue to promote these findings based on your<br /> preprint as well. As stated in our publication, we encourage the use of<br /> low-bias library prep kits such as the Illumina Flex kit.

We noticed that you were wondering which version of<br /> SeqSero2 was used in our study. Information about all programs and versions can<br /> be found in the supplementary material, Table S1 (https://aem.asm.org/content/aem/suppl/2020/02/06/AEM.02265-19.DCSupplemental/AEM.02265-19-s0001.pdf). As<br /> stated, all analyses were performed with SeqSero2 version v.1.0.0, as version<br /> v.1.0.2 was not released before 30th of September 2019 and we<br /> submitted our manuscript to AEM on 2nd of October 2019.

In regard to the nomenclature, we used k-mer mode<br /> and allele-mode as these are the terms used in the official documentation of<br /> SeqSero2 on GitHub (latest version of the readme, current commit: e599e82).

Please do not hesitate to contact me if you have any<br /> further questions.

Best Regards,

Laura Uelze

On 2020-03-30 13:43:11, user Jeremy Green wrote:

Could the authors add illustrations of the chemical structures of the EIDD compounds? These are, after all, the subject of the paper.

On 2020-03-21 20:15:14, user Grant Jacobs wrote:

Could the authors add 'in mice' in the abstract, please? Better still, in the title.

On 2020-03-30 12:26:25, user Afolake wrote:

Waoh! I have been working on this gene for five years now and is amazing and as well mixed feeling to see this paper published before we could get my manuscript out. I hope we will be able to collaborate with your research group on FAM111B proteins for future work. Congrats on this beautiful and thorough work.

On 2020-03-30 12:04:26, user Aaron wrote:

It's a nice paper with an elegant model. My concern is its all based on the assumption that everything controlling their HSF1-reporter (not a HSR reporter directly) affects the HSR. i.e. that HSF1 is the be all and end all. We've shown in bat cells that multiple HSP's are upregulated (i.e. the HSR) in the absence of high HSF1/2 levels meaning there are other regulators of the HSR independent of HSF1/2. Surely this is the case in other cells too and may explain some of their different responses between stressors.

On 2020-03-30 11:46:50, user Nikolas K Haass wrote:

This is an amazingly elegant study showing that p53 plays a central role in lymphedema and can be targeted as a therapeutic strategy.

On 2020-03-30 07:31:03, user L A wrote:

Either there's some problem with Protein E, or there's some new biology

there. In the work it's found to interact with DNA-related proteins. In

Fig 4D some similarity to histones is proposed, but the Protein E

sequence compared (there in the figure to H2A) is incorrect. OK yes even

with the correct sequence there is some similarity, but it looks quite

low and limited just to half of what the authors propose. The match

could just be casual because this is a putative transmembrane segment

(Protein E is probably membrane bound) and the corresponding element in

the H2A sequence is quite hydrophobic.

On 2020-03-29 15:42:26, user Kevin Olivieri wrote:

What a great starting point for tailoring existing treaments to SARS-CoV-2! Are your collaborators exploring synergies with some the approved compounds and the potent antivirals ribavirin and remdesivir? It may allow you to focus only on drugs already in the clinic. If you give remdesivir, which targets NSP12, and ponatinib which targets the NSP12 interactor RIPK1, you might block viral replication more potently than either compound alone. Additionally, impeding viral protein function may enhance the relative effect of ribavirin targets. For example, we know ribavirin inhibits IMPDH1 potently, but the NSP14 interactor IMPDH2 less potently. Since, IMPDH2 interacts with 4 ORF8 interactors, perhaps using drugs that block ORF8s interactants from functioning completely would combine to block IMPDH2 more effectively. Of note, two of these proteins, NPC2 and OS9, are highly expressed in lung.

On 2020-03-28 09:39:12, user Satish wrote:

Thanks. Great work. As a clinician, my understanding is that the virus essentially wipes of ACE2 and allows unchallenged action of ACE in the lung. Obviously antiviral drugs as listed are important in decreasing the Viraemia and should be considered fairly early in the clinical phase, at least in vulnerable population: simple drugs are dihydrochloroquine and Azithromycin ( during febrile phase). Once they are breathless with ARDS, nebulised Losartan or similar drug might have a bigger role. Obviously if we have access to Remdesivir, Tocilizumab,Interferon,Kaletra etc then they are definitely the big guns. Do you think there is a role for Nebulized Losartan and can you study that ? Thanks

On 2020-03-27 14:39:11, user L A wrote:

Either there's some problem with Protein E, or there's some new biology <br /> there. In the work it's found to interact with DNA-related proteins. In <br /> Fig 4D some similarity to histones is proposed, but the Protein E <br /> sequence compared (there in the figure to H2A) is incorrect. OK yes even<br /> with the correct sequence there is some similarity, but it looks quite <br /> low and limited just to half of what the authors propose. The match <br /> could just be casual because this is a putative transmembrane segment <br /> (Protein E is probably membrane bound) and the corresponding element in <br /> the H2A sequence is quite hydrophobic.

On 2020-03-26 10:38:29, user L A wrote:

Either there's some problem with Protein E, or there's some new biology there. In the work it's found to interact with DNA-related proteins. In Fig 4D some similarity to histones is proposed, but the Protein E sequence compared (there in the figure to H2A) is incorrect. OK yes even with the correct sequence there is some similarity, but it looks quite low and limited just to half of what the authors propose. The match could just be casual because this is a putative transmembrane segment (Protein E is probably membrane bound) and the corresponding element in the H2A sequence is quite hydrophobic.

On 2020-03-26 10:34:07, user Samuel Lampa wrote:

Thank you very much for sharing these results! Just a small note: The link to github in the methods is wrong (www.github.com/momeara/BioC...:KoLPxWbceXJWmR0GUxFtae5Oeu4 "www.github.com/momeara/BioChemPantry/vignette/COVID19)"), and should be: https://github.com/momeara/...

On 2020-03-24 22:30:21, user Sinai Immunol Review Project wrote:

Summary: Gordon et al cloned, tagged and expressed 26 of the 29 SARS-CoV-2 proteins individually in HEK293T cells and used mass spectrometry to identify protein-protein interactions. They identified 332 viral-host protein-protein interactions. Furthermore, they used these interactions to identify 66 existing drugs known to target host proteins or host pathways (eg SARS-CoV-2 N and Orf8 proteins interact with proteins regulated by the mTOR pathway, so mTOR inhibitors Silmitasertib and Rapamycin are possible drug candidates).

Limitations: The main limitation of the study stems from the reductionist model: overexpression of plasmids encoding individual viral proteins in HEK293T cells. This precludes any interactions between the viral proteins, or the combined effects of multiple proteins on the host, as they are expressed individually. Moreover, HEK293T cells come from primary embryonic kidney and therefore might not reflect how SARS-CoV-2 interacts with its primary target, the lung. However, the authors found that the proteins found to interact with viral proteins in their experiments are enriched in lung tissue compared to HEK293Ts.

Significance: The authors provide a “SARS-CoV-2 interaction map,” which may provide potential hypotheses as to how the virus interacts with the host. Further, they identified existing drugs that could disrupt these host-viral interactions and curb SARS-CoV-2 infection. Although these interactions have not been validated, this paper acts as a valuable resourc

On 2020-03-24 01:10:36, user Brian Osborne wrote:

Where are the sequences of the 28 proteins that they expressed? These experiments can't be reproduced without the sequences. Didn't see these in the paper, or in the supplementary material.

On 2020-03-23 23:29:20, user Anna Maria Niewiadomska wrote:

Regarding ORF10:CUL2. Interesting particularly since there's evidence of other coronaviruses interacting with proteins in the ubiquitination pathway. However, two groups performing direct sequencing of SARS-CoV-2 mRNA have not been able to detect ORF10 transcripts and no ORF10 homologues exist in other coronaviridae. Although you do mention there's no evidence for ORF10 expression in the supplementary material, you may want to consider that the ORF10 results may be an artifact of over-expression and mention in the main text.

On 2020-03-23 12:20:17, user Ben H wrote:

The authors should be a lot more cautious about the utility of their conclusions. Relying on antibiotics for their off target attacking mitochondrial ribosomes is probably not a safe treatment strategy, because it can lead to cardio toxicity

On 2020-03-29 23:12:09, user Tyler A. Elliott wrote:

Was interesting to hear about the classification discordance between different databases, it would be great to have more details on that, especially so that the databases or users of the databases can be aware of this. Also having more details on possible mis-classification from the confusion matrices might be useful in a similar sense. Very impressive results though.

On 2020-03-29 21:05:06, user Jianhai XIANG wrote:

Being a marine biologist, I am very interesting in this paper. That's quite cool for Evo-devo of Marine vs terrestrial questions. As I know that it is firstly report of establishment of a marine nematode model for animal functional genomics. The paper is wrote well and will bring<br /> Litoditis marina as an unique satellite marine model to the wellknown biomedical model nematode C. elegans. I belive this study will underpin ongoing work on animal functional genomics, environmental adaptation and developmental evolution.

On 2020-03-29 14:47:57, user Russian,Italian,German,Belgian wrote:

This paper is still circulating, It was withdrawn. See this link https://www.bioworld.com/ar...

Quote"The research “was very shoddily done,” Bedford said. “The sequence differences are not unique to COVID-19. Closely related [bat] coronaviruses have these chunks as well. They are small motifs used by nature over and over again.”

On 2020-03-29 07:31:19, user Robert George wrote:

with regard to enigmatic Chermuchek people, the text states ''This model fits even when ancient European farmers are included in the outgroups, showning that if the long-distance transfer of West European megalithic cultural traditions to people of the Chemurchek culture that has been suggested in the archaeological literature occurred, it must have been through<br /> spread of ideas rather than through movement of people. '''

However, a possible link entails the documented movement of Eur-farmer ancestry toward steppe (e.g. GAC groups - Mathieson 2018). In turn, Yamnaya groups (putative parental source of Afansievo) have ~ 10% EEF admixture c.f. Progress steppe Eneolithic (Wang et al 2019). Thus, we have more than a diffusion of ideas.

On 2020-03-28 08:34:15, user Robert George wrote:

Thanks for all the samples. Wishing for more data from Neolithic China

On 2020-03-26 01:13:06, user Davidski wrote:

Hello authors,

Please note that, despite your claims, there's currently no direct evidence of Afanasievo-derived ancestry in what is now western China during the Iron Age.

Surely, due to the limitations of qpAdm, especially in fine scale analyses, it's wrong to suggest that qpAdm models can provide such evidence.

In fact, and again despite your claims, the uniparental markers of the Iron Age Shirenzigou nomads contradict the idea that their steppe ancestry was mediated via the Afanasievo people.

On 2020-03-29 00:29:13, user Andrew Schaumberg wrote:

Please an extended abstract of this work published as Schaumberg et al 2020 "Machine learning for real-time search and prediction of disease state to aid pathologist collaboration on social media" at the Pathology Visions 2019 conference of the Digital Pahtology Association https://www.ncbi.nlm.nih.go...<br /> I would humbly add that indeed, authors WC and SJC made equal conributions.

On 2020-03-28 23:08:27, user surt_lab wrote:

Hi all,

A few of us got together to discus this preprint over a virtual journal club, and I've cobbled together a variety of comments below that came out of this discussion. As background, those discussing the article had pretty broad backgrounds, but with research focuses within microbe-plant interactions and plant genetics.

Finding genes underlying plant-microbe interactions is certainly of high interest right now (given that we're smack dab in the middle of the age of agricultural microbiomes), and in general we currently have relatively poor understanding of heritability microbes that interact with plants as part of the phytobiome. This manuscript therefore begins to address a couple of major questions in the field right now, and does so using an approach that's been powerful in identifying genes underlying agricultural traits in plants.

All of those discussing this manuscript found the GWAS results quite compelling, and that it is certainly clear that there are a few different taxa that truly have high heritabililty in the sorghum rhizosphere. We were all convinced by the data supporting a host association locus on chromosome 4, and especially so given the follow up experiment using different suites of genotypes to predict microbiome traits. We did think it would be good to highlight (more forwardly) that this predictive experiment took place in a growth chamber and not in the field because strongly highlighting that this association is maintained under different growth conditions significantly strengthens the case.

However, there are also a few parts of this manuscript that remained either unclear or which we somewhat disagreed with the interpretations:

Fig 2: A few of us found Figure 2 to be confusing (specifically, it took a while to figure out what the 5X and 40X meant). Would be good to lay this out a little more clearly in the legend. We also thought that the authors should provide a bit more context in the manuscript as to why this figure is important (the read depth/heritability discussion is a good one, but it would be good to include more context towards this point in the paper itself).

Fig. 3: We were also a bit confounded by Figure 3 (and the methods underlying the data). How were you able to call OTUs across all three experiments given that different platforms would call OTUs with different numbers (did you reanalyze the data for all three experiments on the same OTU calling pipeline?). We were also curious as to whether calling taxa as ASVs here would give you consistent results and whether that would be useful). Lastly, some of us didn't have a deep enough understanding of the relationships between Sorghum and Maize (specifically that they are closely related) to grasp the implications of similarity of heritability in taxa across these experiments. We think it would significantly improve the paper to include more context as to how to interpret these heritability results and specifically what the predictions might be as you step out to more distant taxa (or as you step inward to more closely related taxa to Sorghum).

Fig. 4: Lastly, the one part of this manuscript that we remain thoroughly unconvinced about is the gene gazing within the locus on chromosome 4. There is really no expectation that a gene for association would solely be expressed in roots (and not other parts of the plant). You can handwave a bit and even come up with situations where the locus that's important for heritability in the rhizosphere is not even expressed in roots (if there are architectural or hormone changes in the plant that have feed forward effects on the root for instance). Therefore, we didn't think the gene expression data was useful in the context of trying to figure out what the gene of interest might be in controlling host heritability. Moreover, there are a bunch of other really interesting genes within this locus other than those prominently mentioned within the manuscript that could definitely affect microbial composition (e.g. Exo70, NDR1-like, RGA2...). Would be good in our opinions to minimize interpretations of the genes underlying the phenotypes without more follow up genetics experiments and crosses.

Overall, really cool paper that moves the agricultural microbiome field forward conceptually. The GWAS data and association is quite compelling, and we're interested to see follow ups to this work that hammers out what the genes truly are that underlie the heritability phenotype and how heritability of these taxa change as you step to more distant plant groups.

On 2020-03-28 22:38:48, user vAsisTha wrote:

Hi, the 1240K genotypes folder on the data site is empty.<br /> link https://edmond.mpdl.mpg.de/...

On 2020-03-28 17:55:44, user Alexey Kovalev wrote:

Dear colleagues,

your work is of great interest to world science, and its results will undoubtedly have a great impact on the future. However, it cannot be published in its current form in connection with the foregoing.<br /> In your article, you give acknowlegments to us: i.e. D. Erdenebaatar and A.Kovalev for contributing archaeological material to study. However, no one asked permission from us, and did not contact us with questions about the attribution of the results of our excavations. In this regard, in your article we found many errors in the attribution of graves, the bones of which were used. I protest that these materials were so unprofessionally used, I believe that you need to clarify the attribution of materials from our excavations.We are ready to help you with this. Prior to this, we consider it impossible to publish this article.

In addition: you didn't use archaeological numbering of samples. Numbers like AT_000 are numbers given by antropologists of National University. Not all these numbers we can check and it need a some work to compare these numbers with real archaeological sites. All excavated barrows have names and numbering of cemeteries (sites), numbering of excavated kurgans and graves and numbering of burials if more than one burial situated in kurgan. Some burials might be secondary and belong to more later periods.

I can inform you firstly only about some mistakes about barrows excavated by us and used by you without our permission.<br /> 1. Sample KUR001 (AT_635) dasn't belong to Afanasievo. It was rectangular barrow in Chemurchek tradition (named Kurgak govi #2) situated nearby Afanasievo kurgan (named Kurgak govi #1). This sample would need to be combined with sample KUM001 AT_628 (Kumdi govi #1 burial 2). It are secondary burials in ritual structures of Chemurchek types. I think now it was mixed group of people with Chemurchek and East Kazakhstan cultural traditions.<br /> 2. Sample SBG001 (AT_960) dasn't belong to Munkhkhairkhan culture. This barrow named as Shar gov' 3 excavated by me and Munkhbayar oin Bayan-Ulgii aimag, Tsengel sum in 2014. Construction of this kurgan and burial custom is wery similar to Begazy-Dandybai culture (Late Andronovo) of neighbouring Kazakhstan. This burial dated back to more later period as Munkhkhairkhan culture (to 12 cent BCE). It will be very interesting to compare it with Kazakhstan LBA.

1. Sample ULI003 (AT_680) dasn't belong to Baitag culture. It was burial Uliastai dood denzh, kurgan #5, burial on the former surface in stone cist on the back in stretched position. Newly discovered Baitag culture characterized by burials in small pits in position with knees up. This kurgan belong to same cultural group like ULZ001 (AT_674) (ULastai Zastava (Khudzhirtyn gol) 1, kurgan #2) with similar consrtruction and burial custom. AT_674 mistakely put in spreadsheed yo Sukhbaatar aimag but it is neigbouring cemetery in the South of Khovd aimag.

To Baitag culture really belong only sample ULI004 (AT_672). This burial #7 was excavated by me in Uliastai dund denzh cemetery of newly dicovered Baitag culture!

1. Sample ULI001 (AT_676) have archaeological name Ulastai Zastava (Khudzhirtyn gol) 2, kurgan 1, burial #4 (main burial). This grave belong to Mongun-Taiga type of LBA graves in Mongolia and Tuva (stretched on side in narrow pit covered by stone slabs). It is not clear by now what cultural groups there were in Western Mongolia in this time and this sample need to combine with other LBA samples named in your article incorrectly DSKC (it is not correect name because Mongun-Taiga type of kurgans never connected with deer stones etc.). As regard as "Sagsai" culture we never use this name because Ts.Tubat didn't gave definition of it.

Yours sincerely,

Alexey Kovalev,

Institute of Archaeology, Russian Academy of Sciences

E-mail: chemurchek@mail.ru

On 2020-03-26 20:03:35, user William J Warman wrote:

New preprint. A dynamic 6,000-year genetic history of Eurasia's Eastern Steppe.

On 2020-03-28 22:14:30, user Houcemeddine Othman wrote:

Good to know that our conclusions are corroborated by your pre-print https://doi.org/10.1101/202...

On 2020-03-19 20:31:32, user Aaron wrote:

The claims made here should be tempered by the fact that these are strictly analytical models. Until the appropriate benchwork can be completed, the authors should be cautious in claiming proof of enhanced infectivity. Further, the pangolin data, while suggestive of potential infection, doesn't take into account other differences in the virus.

On 2020-03-28 20:38:10, user Sinai Immunol Review Project wrote:

Summary: Using a transgenic human Angiotensin-converting enzyme 2 (hACE2) mouse that has previously been shown susceptible to infection by SARS-CoV, Bao et al. create a model of pandemic 2019-nCoV strain coronavirus. The model includes interstitial hyperplasia in lung tissue, moderate inflammation in bronchioles and blood vessels, and histology consistent with viral pneumonia at 3 days post infection. Wildtype did not experience these symptoms. In addition, viral antigen and hACE2 receptor were found to co-localize the lung by immunofluorescence 3-10 days post infection only in the hACE2 infected mice.

Limitations: The characterization of the infection remains incomplete, as well as lacking characterization of the immune response other than the presence of a single antiviral antibody. Though they claim to fulfill Koch’s postulates, they only isolate the virus and re-infect Vero cells, rather than naive mice.

Significance: This paper establishes a murine model for 2019-nCoV infection with symptoms consistent with viral pneumonia. Though not fully characterized, this model allows in vivo analysis of viral entry and pathology that is important for the development of vaccines and antiviral therapeutics.

On 2020-03-28 20:13:13, user Sinai Immunol Review Project wrote:

Title: <br /> SARS-CoV-2 and SARS-CoV Spike-RBD Structure and Receptor Binding Comparison and Potential Implications on Neutralizing Antibody and Vaccine Development<br /> The main finding of the article: <br /> This study compared the structure of SARS-CoV and SARS-CoV-2 Spike (S) protein receptor binding domain (RBD) and interactions with ACE2 using computational modeling, and interrogated cross-reactivity and cross-neutralization of SARS-CoV-2 by antibodies against SARS-CoV. While SARS-CoV and SARS-CoV-2 have over 70 % sequence homology and share the same human receptor ACE2, the receptor binding motif (RBM) is only 50% homologous.<br /> Computational prediction of the SARS-CoV-2 and ACE2 interactions based on the previous crystal structure data of SARS-CoV, and measurement of binding affinities against human ACE2 using recombinant SARS-CoV and SARS-CoV-2 S1 peptides, demonstrated similar binding of the two S1 peptides to ACE2, explaining the similar transmissibility of SARS-CoV and SARS-CoV-2 and consistent with previous data (Wall et al Cell 2020).<br /> The neutralization activity of SARS-CoV-specific rabbit polyclonal antibodies were about two-order of magnitude less efficient to neutralize SARS-CoV-2 than SARS-CoV, and four potently neutralizing monoclonal antibodies against SARS-CoV had poor binding and neutralizing activity against SARS-CoV-2. In contrast, 3 poor SARS-CoV-binding monoclonal antibodies show some efficiency to bind and neutralize SARS-CoV-2. The results suggest that that antibodies to more conserved regions outside the RBM motif might possess better cross-protective neutralizing activities between two strains.<br /> Critical analysis of the study: <br /> It would have been helpful to show the epitopes recognized by the monoclonal antibodies tested on both SARS-CoV, SARS-CoV-2 to be able to make predictions for induction of broadly neutralizing antibodies. The data on monoclonal antibody competition with ACE2 for binding to SARS-CoV RBD should have also included binding on SARS-CoV2, especially for the three monoclonal antibodies that showed neutralization activity for SARS-CoV2. Because of the less homology in RBM sequences between viruses, it still may be possible that these antibodies would recognize the ACE2 RBD in SARS-CoV-2.<br /> The importance and implications for the current epidemics:<br /> It is noteworthy that immunization to mice and rabbit with SARS-CoV S1 or RBD protein could induce monoclonal antibodies to cross-bind and cross-neutralize SARS-CoV-2 even if they are not ACE2-blocking. If these types of antibodies could be found in human survivors or in the asymptomatic populations as well, it might suggest that exposure to previous Coronavirus strains could have induced cross-neutralizing antibodies and resulted in the protection from severe symptoms in some cases of SARS-CoV2.

On 2020-03-28 20:09:30, user Sinai Immunol Review Project wrote:

Title: <br /> Reinfection could not occur in SARS-2 CoV-2 infected rhesus macaques<br /> The main finding of the article: <br /> This study addresses the issue or acquired immunity after a primary COVID-19 infection in rhesus monkeys. Four Chinese rhesus macaques were intratracheally infected with SARS-CoV-2 and two out of the four were re-infected at 28 days post initial infection (dpi) with the same viral dose after confirming the recovery by the absence of clinical symptoms, radiological abnormalities and viral detection (2 negative RT-PCR tests). While the initial infection led the viral loads in nasal and pharyngeal swabs that reach approximately 6.5 log10 RNA copies/ml at 3 dpi in all four monkeys, viral loads in the swabs tested negative after reinfection in the two reinfected monkeys. In addition, the necropsies from a monkey (M1) at 7 days after primary infection, and another monkey (M3) at 5 days post reinfection, revealed the histopathological damages and viral replication in the examined tissues from M1, while no viral replication as well as no histological damages were detected in the tissues from M3. Furthermore, sera from three monkeys at 21 and 28 dpi exhibited neutralizing activity against SARS-CoV-2 in vitro, suggesting the production of protective neutralizing antibodies in these monkeys. Overall, this study indicates that primary infection with SARS-CoV-2 may protect from subsequent exposure to the same virus.<br /> Critical analysis of the study: <br /> In human, virus has been detected by nasopharyngeal swabs until 9 to 15 days after the onset of symptoms. In the infected monkeys in this study, virus were detected from day 1 after the infection, declining to undetectable level by day 15 post infection. It may suggest that there is a faster viral clearance mechanism in monkeys, therefore the conclusions of reinfection protection for humans need to be carefully considered. In addition, only two monkeys were re-infected in this study and the clinical signs of these monkeys were not similar: M3 did not show weight loss and M4 showed relatively higher fever on the day of infection and the day of re-challenge. <br /> The importance and implications for the current epidemics:<br /> This study showed clear viral clearance and no indications of relapse or viremia after a secondary infection with SARS-CoV-2 in a Chinese rhesus macaque model. These results support the idea that patients with full recovery (two negative RT-PCR results) may also be protected from secondary SARS-CoV-2 infection. Recovered patients may be able to reintegrate to normal public life and provide protective serum perhaps even if having had a mild infection. The results are also encouraging for successful vaccine development against SARS-CoV-2.

On 2020-03-19 21:12:02, user Brindusa wrote:

yes, we need studies in humans but the results are encouraging. We share many similarities with these monkeys (we share 93% of their DNA).

On 2020-03-18 18:38:34, user piper m treuting wrote:

What were the sexes of the Rhesus?

On 2020-03-17 08:06:36, user Adam pearce wrote:

Thank you for this research and very encouraging data. I've just got back form Tesco so nice to be back to the real world again. Its seems to be increasingly likely that us humans will develop immunity. We will get through this thanks to people like Linlin and colleagues alike.

On 2020-03-15 23:03:16, user John Sullivan wrote:

Interesting report and wondering if "protection" after re-challenge is T-cell or B-cell mediated?<br /> Still so little know about what constitutes a good immune response to SARS-CoV-2 infection and especially important to know if "protection" is antibody-mediated as it provide a possible treatment option using a hyperimmune IVIG from recovered COVID-19 patients.

On 2020-03-15 17:25:33, user Andy wrote:

Based on the study design shown in Figure 1, the conclusion that reinfection cannot occur is based on N=2 monkeys. And 1 of those 2 was euthanized 5 days after re-exposure.

Please read the paper carefully before jumping to conclusions based on 1 monkey.

On 2020-03-15 16:39:44, user joncloke wrote:

My knowledge of epidemiology is very slight indeed, however I read a fair bit about the 1918 flu epidemic in which the second wave was deadlier than the first because the virus had evolved.

I know Covid-19 is very different from flu, but I wanted to ask (since we already know it is evolving) if it isn't possible for the same thing to happen with Coronavirus?

Which might mean trying to infect monkeys with the same original strain of virus would be pretty meaningless?

On 2020-03-15 13:33:38, user rich caldwell wrote:

Let's hope this is: a.) accurate and reproducible; and b.) extensible and applicable to homo sapiens.

On 2020-03-15 11:23:20, user Charles Agoti wrote:

....Excellent stuff! Will this be reproducible in humans? How long will immunity against reinfection last ( natural or vaccine)? Which antigens are most important to develop immunity?

On 2020-03-28 17:35:38, user Binbin Chen wrote:

You might be interested in this earlier paper too:<br /> https://www.biorxiv.org/con...

On 2020-03-28 16:02:04, user Eason wrote:

CD147 is most known as the receptor for malaria infecting human blood cells. But, SARS-CoV-2 is rarely detected in the blood from COVID-19 patients. Further study should test SARS-CoV-2 in CD147 positive blood cell from COVID-19 patients.

On 2020-03-26 01:31:49, user Sinai Immunol Review Project wrote:

Summary: <br /> The authors propose a novel mechanism of SARS-CoV-2 viral entry through the interaction of the viral spike protein (SP) and the immunoglobulin superfamily protein CD147 (also known as Basigin). Using an in-house developed humanized antibody against CD147 (maplazumab), they show that blocking CD147 decreases viral replication in Vero E6 cells. Using surface plasmon resonance (SPR), ELISA, and Co-IP assays, they show that the spike protein of SARS-CoV-2 directly interacts with CD147. Lastly, they utilize immune-election microscopy to show spike protein and CD147 localize to viral inclusion bodies of Vero E6 cells.

Critical Analysis: <br /> The authors claim that an anti-CD147 antibody (Meplazumab) inhibits SARS-CoV-2 replication by testing cell growth and viral load in cells infected with SARS-CoV-2, however there are key pieces of this experiment that are missing. First, the authors fail to use a non-specific antibody control. Second, the authors claim that viral replication is inhibited, and that they test this by qPCR, however this data is not shown. To further prove specificity, the authors should introduce CD147 to non-susceptible cells and show that they become permissive. <br /> The authors claim that there is a direct interaction between CD147 and SP through SPR, ELISA, and Co-IP, and this data seems generally convincing. The electron microscopy provides further correlative evidence that SARS-CoV-2 may interact with CD147 as they are both found in the same viral inclusion body. A quantification of this data would make the findings more robust. <br /> Finally, the data in this paper lacks replicates, error bars, and statistics to show that the data are reproducible and statistically significant.

Implications:<br /> It has been shown in various studies that SARS-CoV-2 binds to the cell surface protein ACE2 for cell entry, yet ACE2 is highly expressed in heart, kidney, and intestinal cells, raising the concern that blocking ACE2 would result in harmful side effects. [1] CD147 on the other hand is highly expressed in various tumor types, inflamed tissues, and pathogen infected cells, suggesting that the inhibition of CD147 would not result in major side effects. [2,3] The research in this paper has resulted in an ongoing clinical trial in China to test the safety and efficacy of anti-CD147 Meplazumab to treat COVID-19. (ClinicalTrails.gov identifier NCT04275245).

1. Kosugi T, Maeda K, Sato W, Maruyama S, Kadomatsu K. CD147 (EMMPRIN/Basigin) in kidney diseases: from an inflammation and immune system viewpoint. Nephrol Dial Transplant. 2015; 30: 1097-1103.
2. Su H, Yang Y. The roles of CyPA and CD147 in cardiac remodelling. EXP MOL PATHOL. 2018; 104: 222-226
3. Weidle U, Scheuer W, Eggle D, Klostermann S, Stockinger, H. Cancer-related Issues of CD147. Cancer Genomics and Proteomics. 2010; 7: 157-170

Reviewed as part of a project by students, postdocs, and faculty at the Immunology Institute of the Icahn School of Medicine at Mount Sinai.

On 2020-03-16 07:42:23, user Ahmed Sayed Abdel-Moneim wrote:

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

The authors built their research on the finding of Chen et al., 2005<br /> [https://academic.oup.com/ji...] and the similarity between SARS-CoV-2 SARS-CoV spike protein [S not SP as mentioned by the authors]. <br /> However, Chen et al., 2005, confirmed that SARS-CoV spike protein could not bind to CD147 but they found that the N protein could only bind and speculated that N protein is relocated to the surface from the core during maturation of the virus.

On 2020-03-28 15:34:52, user David BA wrote:

Very good work. If this is still a working paper, could you please expand on why was it expected to obtain better ML results from random rather than GC corrected models? Wouldn't the specificity of the correction maybe help the analysis?

On 2020-03-27 15:47:39, user antonio_j_p_rez wrote:

Machine learning will solve next challenges in bioinformatics. Good work

On 2020-03-27 09:46:55, user Fernando Delgado Chaves wrote:

Great work! Understanding the mechanisms underlying topoisomerase-mediated gene regulation could yield outstanding results in disease treatment.

On 2020-03-28 15:12:41, user Neal Haddaway wrote:

Can you provide more details on which databases and indexes were searched in WoS, please? WoS is not a database (in fact, neither is Google Scholar, it is a search engine). WoS is a platform to access different databases, and the databases available will depend on subscriptions. Within WoS Core Collections, the indexes subscribed to differ across institutions as well. If you just searched WoS from different locations, the results would of course differ by design - different institutions have different sets of bibliographic information available.

On 2020-03-28 14:49:10, user Sinai Immunol Review Project wrote:

Main findings: This study employs a series of bioinformatic pipelines to identify T and B cell epitopes on spike (S) protein of SARS-CoV-2 and assess their properties for vaccine potential. To identify B cell epitopes, they assessed structural accessibility, hydrophilicity, and beta-turn and flexibility which are all factors that promote their targeting by antibodies. To identify T cell epitopes, they filtered for peptides with high antigenicity score and capacity to bind 3 or more MHC alleles. Using the protein digest server, they also demonstrated that their identified T and B cell epitopes are stable, having multiple non-digesting enzymes per epitope. Epitopes were also determined to be non-allergenic and non-toxin as assessed by Allergen FP 1.0 and ToxinPred, respectively. For T cell epitopes, they assessed the strength of epitope-HLA interaction via PepSite. Overall, they predict four B cell and eleven T cell epitopes (two MHC I and nine MHC II binding) to pass stringent computational thresholds as candidates for vaccine development. Furthermore, they performed sequence alignment between all identified SARS-CoV-2 S protein mutations and predicted epitopes, and showed that the epitopes are conserved across 134 isolates from 38 locations worldwide. However, they report that these conserved epitopes may soon become obsolete given the known mutation rate of related SARS-CoV is estimated to be 4x10-4/site/year, underscoring the urgency of anti-viral vaccine development.

Limitations of the Study: While spike (S) protein may have a critical role in viral entry into host cells and their epitope prediction criterion were comprehensive, this study did not examine other candidate SARS-CoV-2 proteins. This point is particularly important given that a single epitope may not be sufficient to induce robust immune memory, and recent approaches involve multi-epitope vaccine design. Furthermore, their study only included a direct implementation of various published methods, but did not validate individual bioinformatic tools with controls to demonstrate robustness. Finally, it is critical that these predicted epitopes are experimentally validated before any conclusions can be drawn about their potential as vaccine candidates or their clinical efficacy.

Relevance: This study provides a computational framework to rapidly identify epitopes that may serve as potential vaccine candidates for treating SARS-CoV-2.

On 2020-03-28 13:18:15, user Euan Pyle wrote:

Hi, this is nice work! I noticed at one point you say "Thus, lipid-protein interactions are likely crucial in the activity of the majority of membrane proteins but actual functional data to support this hypothesis is scarce." which is true! Although, we published a few papers finding a functional role for protein-lipid interactions if you are convinced by the argument we present:

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

https://pubs.acs.org/doi/ab...

On 2020-03-28 12:23:17, user xyyman wrote:

I am surprised so maybe you can explain. There are no mtDNA H1,H2 and H3 subclade found, This would imply no western European genetic influence? Your haplogroup data shows more African influence than Western European influence doesn’t it? mtDNA L and M(Kivilsid et al) having undoubtedly African origin. So why use the statement “*possibly* reflecting the into-Africa influx of human migration.”? Furthermore isn’t the ancestral clades of mtDNA H found in Sudan and of course Yemen. Regions south of Egypt? In fact looking at the phyloTree of R. R shows more diversity in Sudan rather than Yemen. Correct me if I am wrong?

On 2020-03-28 01:05:18, user Sinai Immunol Review Project wrote:

Summary: Given the sequence similarity of the surface spike glycoprotein (S) of SARS-CoV-2 and SAR-CoV, Yuan et al. (2020) propose that neutralizing antibodies isolated from convalescent SARS-CoV patients may offer insight into cross-reactive antibodies targeting SARS-CoV-2. In particular, they find that the receptor-binding domain (RBD) of SARS-CoV-2 S protein shares 86% sequence similarity with the RBD of SARS-CoV S protein that binds to the CR3022 neutralizing antibody. CR3022 also displays increased affinity for the “up” conformation of the SARS-CoV-2 S protein compared to the “down” conformation as it does for the SARS-CoV S protein. Therefore, the authors propose that this cross-reactive antibody may confer some degree of protection in vivo even if it fails to neutralize in vitro.

Limitations: Although the authors offer a logical rationale for identifying cross-reactive neutralizing antibodies derived from SARS-CoV, their study using only CR3022 failed to demonstrate whether this approach will be successful. After all, CR3022 failed to neutralize in vitro despite the binding affinity to a similar epitope on SARS-CoV-2. They would benefit from testing more candidates and using an in vivo model to demonstrate their claim that protection may be possible in the absence neutralization if combinations are used in vivo.

Significance: The ability to make use of previously characterized neutralizing antibodies for conserved epitopes can expedite drug design and treatment options.

On 2020-03-27 19:42:22, user Sinai Immunol Review Project wrote:

This study reconsiders the use of inhaled corticosteroids in the treatment of pneumonia by coronavirus. Corticosteroids were associated with increased mortality for SARS in 2003 and for MERS in 2013, probably due to that fact that systemic corticosteroids suppress the innate immune system, resulting in increased viral replication. However, some steroid compounds might block coronavirus replication. The authors screened steroids from a chemical library and assessed the viral growth suppression and drug cytotoxicity. Ciclesonide demonstrated low cytotoxicity and potent suppression of MERS-CoV viral growth. The commonly used systemic steroids cortisone, prednisolone and dexamethasone did not suppress viral growth, nor did the commonly used inhaled steroid fluticasone. To identify the drug target of virus replication, the authors conducted 11 consecutive MERS-CoV passages in the presence of ciclesonide or mometasone, and they could generate a mutant virus that developed resistance to ciclesonide, but not to mometasone. Afterwards, they performed next-generation sequencing and identified an amino acid substitution in nonstructural protein 15 (NSP15) as the predicted mechanism for viral resistance to ciclesonide. The authors were able to successfully generate a recombinant virus carrying that amino acid substitution, which overcome the antiviral effect of ciclesonide, suggesting that ciclosenide interacts with NSP15. The mutant virus was inhibited by mometasone, suggesting that the antiviral target of mometasone is different from that of ciclesonide. Lastly, the effects of ciclesonide and mometason on suppressing the replication of SARS-CoV-2 were evaluated. Both compounds were found to suppress viral replication with a similar efficacy to lopinavir.

Limitations of Analysis:<br /> Most of the experiments, including the identification of the mutation in NSP15 were conducted with MERS-CoV. This is not the closest related virus to SARS-CoV-2, as that would be SARS-CoV. Thus, to repeat the initial experiments with SARS-CoV, or preferably SARS-CoV-2, is essential. The manuscript should address this and, therefore, it will require considerable editing for organization and clarity. Also, in terms of cell immunogenic epitopes, while SARS-CoV-2 spike protein contains several predicted B and T cell immunogenic epitopes that are shared with other coronaviruses, some studies have shown critical differences between MERS-CoV, SARS-CoV and SARS-CoV-2. A main criticism is that the authors only used VeroE6/TMPRSS2 cells to gauge the direct cytotoxic effects of viral replication. To evaluate this in other cell lines, including human airway epithelial cells, is crucial, as the infectivity of coronavirus strains greatly varies in different cell lines,

Nevertheless, these findings encourage evaluating ciclesonide and mometasone as better options for patients with COVID-19 in need of inhaled steroids, especially as an alternative to other corticosteroids that have been shown to increase viral replication in vitro. This should be evaluated in future clinical studies.

This review was undertaken by Alvaro Moreira, MD as part of a project by students, postdocs and faculty at the Immunology Institute of the Icahn School of Medicine, Mount Sinai.

On 2020-03-22 05:59:23, user Jortiz3 wrote:

Interesting. Could SARS-CoV-2 generate a similar resistant mutation MERS did? Is that a concern as a candidate drug? Or am I misunderstanding.

On 2020-03-27 18:20:49, user Alexander Jaffe wrote:

As I said on twitter, very interesting work! I had a quick question/comment for you about the metabolic association of the new rubisco sequences related to Form III-a:

While it is certainly possible that they are associated with the nucleotide scavenging pathway described for some archaea, it may also be worth exploring whether your sequences could be associated with the RHP, as for at least one methanogen encoding the III-a rubisco: (https://www.nature.com/arti.... Did you get a chance to look for PRK in the reported genomes, or the AMP phosphorylase and isomerase associated with the nucleotide scavenging pathway?

In either case, an exciting opportunity to potentially validate or expand on these hypotheses using the transcriptomes!

On 2020-03-27 16:56:04, user Nathan Meyers wrote:

The y-axis expression values on ACE2 and TMPRSS2 expression are extremely low. As a liver biologist, I don't think studying SARS-CoV-2 in the liver is very relevant and these data support my thoughts. This manuscript would also benefit from a more thorough exploration of liver progenitor markers. SOX9 is also expressed in a subset of hepatocytes

On 2020-03-27 16:45:50, user Sinai Immunol Review Project wrote:

This study harnesses bioinformatic profiling to predict the potential of COV2 viral proteins to be presented on MHC I and II and to form linear B-cell epitopes. These estimates suggest a T-cell antigenic profile distinct from SARS-CoV or MERS-CoV, identify focused regions of the virus with a high density of predicted epitopes, and provide preliminary evidence for adaptive immune pressure in the genetic evolution of the virus.

While the study performs a comprehensive analysis of potential epitopes within the virus genome, the analysis relies solely on bioinformatic prediction to examine MHC binding affinity and B-cell epitope potential and does not capture the immunogenicity or recognition of these epitopes. Future experimental validation in data from patients infected with SARS-CoV-2 will be important to validate and refine these findings. Thus some of the potential conclusions stated, including viral evolution toward lower immunogenicity or a dominant role for CD4+ T-cells rather than CD8+ T-cells in viral clearance, require further validation.

In the future, these findings may help direct peptide vaccine design toward relevant epitopes and provide intriguing evidence of viral evolution in response to immune pressure.

On 2020-03-27 14:57:56, user Nikolaos Sgourakis wrote:

3D pMHC models available at: https://rosettamhc.chemistr...

On 2020-03-27 14:13:43, user haiguang liu wrote:

For those who want to use the software online, there is a server http://liulab.csrc.ac.cn/de...

Source code is available on Github: https://github.com/LiuLab-C...

On 2020-03-27 12:42:55, user UAB Journal Club wrote:

Review of Dhasmana et al. “Bacillus anthracis chain length, a virulence determinant, is regulated by a transmembrane Ser/Thr kinase”<br /> University of Alabama at Birmingham Bacterial Pathogenesis and Physiology Journal Club

Summary<br /> B. anthracis chain length is an important virulence determinant that is associated with blockage of lung capillaries, leading to hypoxia and lung tissue damage in a murine model of infection. The authors identify and characterize a serine/threonine kinase PrkC as a regulator of bacterial chain length and cell division. A prkC knockout was found to contain shorter chain lengths than wildtype B. anthracis. In this mutant, they observed upregulation of BslO, a murein hydrolase which catalyzes daughter cell formation, and S-layer protein Sap, which is needed for BslO localization at the septal region and its subsequent murein hydrolysis activity. A decrease in cell wall thickness and increase in multi-septa was also observed in the prkC mutant.

Overall, we have found this paper to be well-written, with the results and methods clearly described and the authors’ conclusions convincing. With that said, we have some comments that may be beneficial to address and some additional questions for the authors.

Major Comments<br /> 1. The authors make fairly strong conclusions about the differences in the growth curves between the wild-type and prkC mutant strain. However, chain length has a major impact on OD600 (see Stevenson et al. 2016 Sci Rep 6:38828), which is not taken into account here. It would be useful to calibrate the CFU / OD ratio for each of these strains. The Methods do not mention any dilution of the samples to measure OD, but I presume this must have been done to obtain accurate measurements of OD600 > 1. Please specify.

1. It is clear from the Introduction and Discussion that PrkC is a known virulence determinant and regulator in B. anthracis, but its known regulon and signaling roles are not thoroughly described. The authors should replace some of the introduction of virulence mechanisms in B. anthracis (which is not really relevant to this paper) with an introduction to what is currently known about the signals PrkC responds to and how it controls gene and protein expression.

2. Some explanations in the results section are speculative and would benefit from

experiments to substantiate the claims be shifted to the discussion section.<br /> A. “In the presence of PrkC, synthesis of these molecule(s) is probably downregulated to allow bacteria to grow as chains.”

B. “Altogether, these results indicate that during bacterial growth, PrkC maintains an optimum level of BslO, Sap, and EA1 to maintain the chaining phenotype.”

C. “Our initial results suggest that ftsZ is constantly upregulated in the prkC disruption strain. This probably formed the reason for an increase in multi-septa formation observed in the prkC disruption strain possibly due to the mislocalization of FtsZ.”

D. It is noted that there is chaining observed in the prkC deletion mutant in the lag phase and that this may be due to a lack of de-chaining proteins at this time, but this is not further investigated.

i. Is it possible that PrkC is not the only sensing machinery for the chaining phenotype? If PrkC is the only sensing protein regulating the chaining phenotype, one would expect to not see the chaining phenotype at all. The authors claim this could be due to reduced expression or localization of de-chaining proteins at the early timepoints. However, the cultures used for these images were started from an overnight culture where the proteins would theoretically have been expressed. To shift from one phenotype to another from the overnight to a fresh culture would require some form of environmental sensing not provided by PrkC. Which transcriptional regulators does PrkC ultimately regulate? Are one of those transcription factors allosterically regulated by environmental cues, like nutrient starvation, or are any of them regulators of sap?

1. The chain length is highlighted in Figure 1 but never objectively quantified. This is done later (Figure 3) but is arguably more important in these initial observations as they set the foundation for all future conclusions in the paper.

2. The authors note an inconsistency with PrkC growth curve (noted in results section pertaining to Figure 5) to those which are published. One publication cited is from the lab which sourced the mutant strain. This poses concern for strain genetic drift and should be addressed.

3. One of the major arguments is that PrkC is an environmental sensor (“Through this work, we propose that PrkC, a transmembrane kinase with a sensor domain, perceives growth permissive signals and maintains the levels of the primary proteins involved in de-chaining to regulate the chaining phenotype.”). While the results in this manuscript support the claim that PrkC is the most likely Ser/Thr kinase to be involved in this phenotype due to its localization, there is not enough evidence to support the claim that PrkC senses the environment.

4. At the 2-hour timepoint, long chains are observed in the WT and PrkC KO in Figure 6. It seems to be a main point of the manuscript that this should not be the case. It is unclear and not addressed why both are chained in this figure/timepoint.

Minor Comments<br /> 1. Comments regarding visual display of results<br /> a. Figures 2 & 5: bacterial growth curves should be plotted on a semi-log scale. See https://schaechter.asmblog....

b. Figure S1 has unreadably small text.

c. Change the colors on the graphs so that readers with red/green color blindness can distinguish between datasets.

d. Figure legends contain too much of the methods. Move methods to the method section and make legends more concise.

e. Supplemental tables of P values are not necessary.

1. Comments regarding statistical analysis<br /> a. Figure 3: A one-way ANOVA may be more appropriate here; unsure why a two-way ANOVA was used. Additionally, post-hoc analyses for ANOVA are not listed. Please clarify.

b. Figure 4 needs statistics to indicate at which time points each protein level is different between the wild-type and prkC strains. (Similar to what is present in Figure 3.)

1. Comments regarding wording of text<br /> a. The first paragraph of the results section can be included in the introduction. This information distracts from the data and results.

b. The authors refer to the ΔprkC strain as a “disruption” strain. However, from the reference (Shah et al 2008), the strain is a resolved excision construct. It would be more accurate, therefore, to call it a deletion strain, as disruption suggests that there could be a fragment of the original gene present that could be producing a semi-functional peptide.

c. There are multiple instances where unnecessary information could be removed from the text. For instance, the last sentence of the third paragraph of the results section (“In a study on PknB, a membrane-localized PASTA kinase from Mycobacterium tuberculosis, depletion, or over-expression of the kinase was shown to have a significant effect on bacterial morphology leading to cell death.”) does not contribute, at least in the way it is currently written, to the argument at large and could be removed without detracting from the value of the manuscript.

d. Line 64: “Bacterial chaining has been shown to contribute significantly to virulence” would be a more succinct way to describe this.

e. The authors say that they measured expression levels of BslO, Sap, and EA1 at the “indicated time points” and reference Figure 2A and 2B. There are no indicators of what time points in those figures was used. Most of the time points are used but some are not (12 hours, for instance).

Future-Specific Comments<br /> Figure 1B: CV staining to see how many cells per chain would be insightful. Hard to tell if this is a chaining defect or division defect based on this figure.

Figure 1C: The resolution between the WT and ΔprkC strain is different. The WT strain is much cleaner than the knockout strain, however, this is a minor issue and the size bar indicates that the same scale is used.

Figures 2A&B: These images should be overlaid on the same graph for comparison sake. In the section “prkC disruption results in decreased cell wall width and cell septa thickness and increased multi-septa formation”, they even observe that the two curves are not superimposable, meaning that they do not overlap. It would also be beneficial to see the graph of the average chain length in Figure 2 instead of as a separate figure.

Figure 4: If comparisons are going to be made between WT and KO strain expression of Sap and BslO, they should be run on the same gel. If the comparisons are focused on the time course across one strain, it makes more sense to run them as they are, but the focus is on the difference between strains

On 2020-03-27 11:53:58, user Rui F. Oliveira wrote:

Dear Alex and co-authors,<br /> Congratulations on a very interesting study.<br /> We have recently discussed your preprint on our lab journal club and there were two issues that we would like to have your views on:<br /> 1. According to the description of the selection procedure [l.97-100: "Twenty-six females from the four top-ranked groups in each line were then paired with unsorted males to breed the next generation of polarization-selected fish. To establish control lines (n = 3), we took 26 randomly selected females from the remaining groups, and bred from those fish], it looks like your control lines do not represent either random-mating or average population lines but rather anti-polarisation selected lines, since your are removing the high polarisation females from them. If this is the case, they do not represent a traditional control line, but rather another divergent line;<br /> 2. The concern expressed above is apparently confirmed in the Fig S1 where it can be seen that the significant result in replicate 3 is due to a decrease in polarisation of the "control" line rather than to an increase in the polarisation line (i.e. in replicate 3 polarisation in F0 seems similar to F3_P and higher than F3_C), and that in the other two replicates the effect may result from both increases in the polarisation lines (i.e. F3_P > F0) and decreases in the control lines (i.e. F3_C < F0). To make this point more clear maybe it would help to statistically compare differences in polarization between F3 and F0 in polarization and control lines for each replicate.<br /> With compliments on behalf of the Oliveira Lab, <br /> Rui Oliveira <br /> (Integrative Behavioural Biology Group, Gulbenkian Institute of Science, Portugal)

On 2020-03-27 10:01:00, user Amos Bairoch wrote:

Very nice paper. You can add in your final version that your new A549 DHODH-/- cell line will be assigned the RRID: CVCL_YZ46 in the Cellosaurus.

On 2020-03-27 05:12:57, user Steve Lilac wrote:

EMT as a mechanism of trastuzumab and lapatinib resistance. In my opinion the question why in breast cancer epithelial cells are mostly HER2-high and mesenchymal cells are HER2-low is deliberated and perfectly investigated. Answer: because of different chromatin architecture. The paper is suggesting that HER2 gene stands with epithelial phenotype and can be silenced during EMT similar to other epithelial marks. When an epithelial HER2+ cell undergoes EMT, the cell looses HER2 expression due to chromatin closure . Quite sensible. The most parts of study is done on genomics and epigenomics data from depositories which is impressive. I think this is a great example for how different raw data are to be analyzed in a correct way by other researchers to compile meaningful results. The in silico data contains those from patients and cell culture that is confirmed by experimental results.

On 2020-03-26 03:13:16, user Saeed Bahrami wrote:

Excellent work

On 2020-03-25 11:34:19, user Kasra wrote:

Very nice and important finding

On 2020-03-27 01:51:05, user Sinai Immunol Review Project wrote:

Title: Respiratory disease and virus shedding in rhesus macaques inoculated with SARS-CoV-2

Immunology keywords: animal model, pulmonary infiltrates, dynamic of antibody response, cytokine

Summary: Inoculation of 8 Resus macaques with SARS-Cov-2 , which all showed clinical signs of infection (respiratory pattern, reduced appetite, weight loss, elevated body temperature) resulting in moderate, transient disease. Four animals were euthanized at 3dpi, the 4 others at 21 dpi. Study of viral loads in different organs showed that nose swab and throat swabs were the most sensitive, with bronchio-alveolar lavage. Interstitial pneumonia was visible in radiographies and at the histological scale too. Clinically, the macaques had similar symptoms as described in human patients with moderate disease.<br /> Viral shedding was consistently detected in nose swabs and throat swabs immediately after infection but less consistent thereafter which could reflect virus administration route (intranasal, oral). Bronchoalveolar lavages performed as a measure of virus replication in the lower respiratory tract on animals maintained for 21 days, contained high viral loads in 1 and 3dpi. The majority of the animals exhibited pulmonary edema and mild to moderate interstitial pneumonia on terminal bronchioles. In addition to the lung, viral RNA could also be detected throughout the respiratory track where viral replication mainly occurred. <br /> Immunologic responses included leukocytosis, neutrophilia, monocytosis and lymphopenia in the majority of the animals at 1dpi. Lymphocytes and monocytes re-normalized at 2dpi. Neutrophils declined after 3dpi and through 10dpi after which they started to recover. After infection, serum analysis revealed significant increases in IL1ra, IL6, IL10, IL15, MCP-1, MIP-1b, but quick normalization (3dpi). Antibody response started around 7dpi, and the antibody titers stayed elevated until 21dpi (day of animal euthanasia).

Critical analysis: The macaques were inoculated via a combination of intratracheal, intranasal, ocular and oral routes, which might not reproduce how humans get infected. Maybe this can lead to different dynamics in the host immune response. Also, the authors noted that the seroconversion was not directly followed by a decline in viral loads, as observed in covid19 patients.

Critical analysis and Implications: This work confirms that rhesus macaques can be a good model to study Covid-19, as it has been shown by other groups (10.1101/2020.03.13.990226, 10.1101/2020.03.13.990036, 10.1101/2020.03.17.995639). While these experiments recapitulate moderate COVID-19 in humans, the mode of inoculation via a combination of intratracheal, intranasal, ocular and oral routes, might not reproduce how humans get infected and may lead to different dynamics in the host immune response. For example, the authors noted that the seroconversion was not directly followed by a decline in viral loads, as observed in COVID-19 patients. Therefore, it will be interesting to follow their antibody titers longer and further assess the possibility/effect of reinfection in these macaques. It is essential to be able to understand the dynamic of the disease and associated immune responses, and to work on vaccine development and antiviral drug testing.

Review part of a project of students, postdocs and faculty at the Immunology Institute of the Icahn school of medicine, Mount Sinai.

On 2020-03-23 15:10:29, user Jemi Draw wrote:

Great job, NIAID! Why via a combination of intratracheal, intranasal, ocular and oral routes and not only 1 route?

On 2020-03-27 01:19:33, user benjamin vincent wrote:

We've tried to replicate this analysis from the datasets of Van Allen et al Science 2015, Hugo et al Cell 2016, Riaz et al Cell 2017, and Liu et al Nature Medicine 2019. We analyzed the data by getting FASTQ files where we could (except for the Liu et al study, where we used the gene expression matrix file hosted on GitHub). We aligned with STAR and used Salmon for quantification, followed by upper quartile normalization. We calculated hazard ratios derived from CoxPH regression with overall survival as the response variable. Only the Riaz et al dataset had an HR that did not cross one with a p-value less than 0.05. The combined dataset was not statistically significant by that measure (p-value = 0.168). When analyzed by response class (complete response + partial response versus non-response), no study showed a significant difference in FDCSP expression by response class.

On 2020-03-26 21:52:21, user Daniel Hromada wrote:

In the paper "Pneumonia outbreak associated with a new coronavirus of probable bat origin" peer-reviewed & published in Nature, Zhou et al. explicitely state "no evidence for recombination events was detected in the genome of 2019-nCov".

Unfortunately, Zhou et al. provide very little information about methods which lead them to utter such strong statement less then a month after the virus was officially discovered by authorities.

In Your article, You demonstrate the contrary. If proven right, Your discovery would indicate that Zhou et al. published misleading information which managed to pass through all stages of Nature's peer-reviewing process. That's quite serious.

On 2020-03-15 23:34:34, user Kenny Day wrote:

Trying to find pangolin Cov genome sequences. These don’t appear to have been submitted. Definitely not in NCBI.

On 2020-03-26 21:08:36, user Sinai Immunol Review Project wrote:

Main findings<br /> The authors performed long read RNA sequencing using an Oxford Nanopore MinION as well as tandem mass spec (MS) on Vero cells (a cell line derived from kidney cells of the African green monkey that is deficient in interferon) infected with SARS-CoV-2.

The authors found that passage of the virus in Vero cells gave rise to a spontaneous 9 amino acid deletion (679-NSPRRARSV-687 to I) in the spike (S) protein. The deleted sequence overlaps a predicted furin cleavage site at the S1 / S2 domain boundary that is present in SARS-CoV-2 but not SARS-CoV or the closely related bat coronavirus RaTG13, which are cleaved at S1 / S2 by other proteases [1]. Furin cleavage sites at similar positions in other viruses have been linked to increased pathogenicity and greater cell tropism [2]. Loss of this site in SARS-CoV-2 has also already been shown to increase viral entry into Vero but not BHK cells (which are also interferon deficient) [3]. The authors therefore make an important contribution in demonstrating that passage in Vero cells may lead to spontaneous loss of a key pathogenicity-conferring element in SARS-CoV-2.

Critical analysis<br /> As the authors note, a similar study posted earlier by Kim et al., which also passaged SARS-CoV-2 in Vero cells, did not identify any loss in the S protein furin cleavage site [4]. It therefore remains to be determined how likely it is that this mutation spontaneously arises. A quantitative investigation using multiple experimental replicas to understand the spontaneous viral mutation rate at this site and elsewhere would be informative. Also, the mechanistic basis for the higher viral fitness conferred by loss of the furin cleavage site in Vero cells – but, evidently, not in vivo in humans, as this site is maintained in all currently sequenced circulating isolates - remains to be understood.

Due to the high base-call error rate of MinION sequencing, the authors’ bioinformatic pipeline required aligning transcripts to a reference to correct sequencing artifacts. This presumably made it difficult or impossible to identify other kinds of mutations, such as single nucleotide substitutions, which may occur even more frequently than the deletions identified in this work. Pairing long read sequencing with higher-accuracy short-read sequencing may be one approach to overcome this issue.

Relevance<br /> As the authors suggest, animal studies using live virus challenge may need to periodically verify the genomic integrity of the virus, or potentially risk unknowingly using a likely less-pathogenic variant of the virus.

More broadly, the results emphasize the complexity and plasticity of the SARS-CoV-2 viral transcriptome and proteome. For example, the authors found multiple versions of transcripts encoding the nucleocapsid (N) protein, each with different small internal deletions, some of which were verified for translation by MS. A number of peptides arising from translation of unexpected rearrangements of transcripts were also detected. Additionally, the authors identified phosphorylation of a number of viral proteins (N, M, ORF 3a, nsp3, nsp9, nsp12 and S). For any cases where these have functional consequences, targeting the kinases responsible could be an avenue for drug development. Understanding the functional consequences of the mutations, transcript variations, and post translational modifications identified in this study will be important future work.

[1] Wrapp, D. et al. Science. doi:10.1126/science.abb2507<br /> [2] Sun et al. doi:10.1128/JVI.00797-10<br /> [3] Walls et al. doi:10.1016/j.cell.2020.02.058<br /> [4] Kim et al. doi:10.1101/2020.03.12.988865.

This review was undertaken as part of a project by students, postdocs and faculty at the Immunology Institute of the Icahn school of medicine, Mount Sinai.

On 2020-03-26 17:38:55, user Sinai Immunol Review Project wrote:

Summary<br /> The authors explore the antigenic differences between SARS-CoV-2 and SARS-CoV by analyzing plasma samples from SARS-CoV-2 (n = 15) and SARS-CoV (n = 7) patients. Cross-reactivity in antibody binding to the spike protein between SARS-CoV-2 and SARS-CoV was found to be common, mostly targeting non-RBD regions in plasma from SARS-CoV-2 patients. Only one SARS-CoV-2 plasma sample was able to cross-neutralize SARS-CoV, with low neutralization activity. No cross-neutralization response was detected in plasma from SARS-CoV patients.

To further investigate the cross-reactivity of antibody responses to SARS-CoV-2 and SARS-CoV, the authors analyzed the antibody response of plasma collected from mice infected or immunized with SARS-CoV-2 or SARS-CoV (n = 5 or 6 per group). Plasma from mice immunized with SARS-CoV-2 displayed cross-reactive responses to SARS-CoV S ectodomain and, to a lesser extent, SARS-CoV RBD. Similarly, plasma from mice immunized with SARS-CoV displayed cross-reactive responses to SARS-CoV-2 S ectodomain. Cross-neutralization activity was not detected in any of the mouse plasma samples.

Potential limitations<br /> The size of each patient cohort is insufficient to accurately determine the frequency of cross-reactivity and cross-neutralization in the current SARS-CoV-2 pandemic. Recruitment of additional patients from a larger range of geographical regions and time points would also enable exploration into the effect of the genetic diversity and evolution of the SARS-CoV-2 virus on cross-reactivity. This work would also benefit from the mapping of specific epitopes for each sample. Future studies may determine whether the non-neutralizing antibody responses can confer in vitro protection or lead to antibody-dependent disease enhancement.

Implications of the findings in the context of current epidemics<br /> The cross-reactive antibody responses to S protein in the majority of SARS-CoV-2 patients is an important consideration for development of serological assays and vaccine development during the current outbreak. The limited extent of cross-neutralization demonstrated in this study indicates that vaccinating to cross-reactive conserved epitopes may have limited efficacy, presenting a key concern for the development of a more universal coronavirus vaccine to address the global health risk of novel coronavirus outbreaks.

This review was undertaken as part of a project by students, postdocs and faculty at the Immunology Institute of the Icahn school of medicine, Mount Sinai.

On 2020-03-26 14:22:19, user Peter Papesch, AIA wrote:

How about combining the information about the covid-19's spike structure with an analysis of the covid-19's lipid membrane to a) identify and b) specifically attack and destroy covid-19? <br /> I can send an article about a material which pierces the biofilm surrounding any cell (as I understand it, its lipid membrane), so being selective about the type of attack seems important. <br /> papesch@mac.com

On 2020-03-24 20:01:45, user Sinai Immunol Review Project wrote:

Main findings<br /> The authors highlight a human angiotensin-converting enzyme 2 (hACE2), as a potential receptor used by the current Severe Acute respiratory syndrome coronavirus-2 (SARS-CoV-2) as a host factor that allows the virus target human cells. This virus-host interaction facilitates the infection of human cells with a high affinity comparable with SARS-CoV. The authors propose this mechanism as a probable explanation of the efficient transmission of SARS-CoV-2 between humans. Besides, Walls and colleagues described SARS-CoV-2 S glycoprotein S by Cryo-EM along with neutralizing polyclonal response against SAR-CoV-2 S from mice immunized with SAR-CoV and blocking SAR-CoV-2 S-mediated entry into VeroE6 infected cells.

Limitations of the study<br /> The SARS-CoV-2 depends on the cell factors ACE2 and TMPRSS2, this last, according to a recent manuscript by Markus Hoffman et al., Cell, 2020. The authors used green monkey (VeroE6) and hamster (BHK) cell lines in the experiments to drive its conclusions to humans; however, it is well known the caucasian colon adenocarcinoma human cell line (CaCo-2), highly express the hACE2 receptor as the TMPRSS2 protease as well. In humans, ACE2 protein is highly expressed in the gastrointestinal tract, which again, makes the CaCo-2 cell line suitable for the following SARS-CoV-2 studies.

Relevance<br /> The results propose a functional receptor used by SARS-CoV-2 to infect humans worldwide and defining two distinct conformations of spike (S) glycoprotein by cryogenic electron microscopy (Cryo-EM). This study might help establish a precedent for initial drug design and treatment of the current global human coronavirus epidemic.

References<br /> 1. Hoffman M. et al. SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor. https://doi.org/10.1016/j.c...

2.- www.proteinatlas.org/ENSG00...<br /> 3. www.proteinatlas.org/ENSG00...

Review by Gustavo Martinez-Delgado as part of a project by students, postdocs and faculty at the Immunology Institute of the Icahn school of medicine, Mount Sinai

On 2020-03-26 14:06:25, user Mina Bizic wrote:

In case you did not see, the paper has been published on 15th Jan 2020. <br /> https://advances.sciencemag...

On 2020-03-26 12:50:53, user RedSiskin wrote:

Would it be possible to give more precise information on origin and locality of the various specimens in the Supplementary Materials, Table S1? In particular, hagenbecki, which has a phenotype (broad white superciliaries, very broad white collar, light crown, light straw yellow coloration of flanks and upper back, small spots on flanks, large apical black wedges on upper back) matching in all respects the ones in the torquatus group, has in Table S1 given as origin "Gobi, Mongolia", with Longitude and Latitude pointing to a place where no subspecies is expected to occur at all. It would be important to make sure the feather samples were indeed hagenbecki, as its belonging to the strauchi-vlangalii group would be quite surprising. Similarly, alaschanicus has in Table S1 a Longitude and Latitude pointing to a place in Tumed Left Banner near Hohhot, not Alxa Left Banner. At this place kaingsuensis is expected to occur. Alaschanicus, however, is rather expected to occur around the Yaoba oasis region near Alxa, west of the Helan Mountain Range. The Longitude/Latitude data for vrangelii reading in Table S1 (101.50, 36.65) are in fact near Xining where typical strauchi occurs, this must be a typo. The distribution of subspecies in the Sichuan Basin is not known (with elegans, suehschanensis, decollatus, and even strauchi-like birds all reported), so it seems that the assignment of suehschanensis in Suining, far away from the Songpan, needs an explanation, are these wild birds? One should also keep in mind that the precise distribution of kiangsuensis, karpowi, and suehschanensis is very unclear to date, as are the subspecies occurring in the Sichuan Basin and the Mountain ranges north of it, and in the Ordos Plateau.

On 2020-03-26 11:16:15, user Chainarong Amornbunchornvej wrote:

We have the ipADMIXTURE package on R CRAN at https://CRAN.R-project.org/....

On 2020-03-26 02:07:24, user Elisabeth Bik wrote:

Figure 4C has a duplication of 2 sets of panels (far-left column vs far-right column). The arrows are at different positions, though.

I cannot upload an image (although I am logged in through Disqus) but I will refer to my tweet here: https://twitter.com/Microbi...

On 2020-03-26 01:04:02, user Sinai Immunol Review Project wrote:

Using in silico bioinformatic tools, this study identified putative antigenic B-cell epitopes and HLA restricted T-cell epitopes from the spike, envelope and membrane proteins of SARS-CoV-2, based on the genome sequence available on the NCBI database. T cell epitopes were selected based on predicted affinity for the more common HLA-I alleles in the Chinese population. Subsequently, the authors designed vaccine peptides by bridging selected B-cell epitopes and adjacent T-cell epitopes. Vaccine peptides containing only T-cell epitopes were also generated.<br /> From 61 predicted B-cell epitopes, only 19 were exposed on the surface of the virion and had a high antigenicity score. A total of 499 T-cell epitopes were predicted. Based on the 19 B-cell epitopes and their 121 adjacent T-cell epitopes, 17 candidate vaccine peptides were designed. Additionally, another 102 vaccine peptides containing T-cell epitopes only were generated. Based on the epitope counts and HLA score, 13 of those were selected. Thus, a total of 30 peptide vaccine candidates were designed.

While this study provides candidates for the development of vaccines against SARS-CoV-2, in vitro and in vivo trials are required to validate the immunogenicity of the selected B and T cell epitopes. This could be done using serum and cells from CoV-2-exposed individuals, and in preclinical studies. The implication of this study for the current epidemic are thus limited. Nevertheless, further research on this field is greatly needed.

This review was undertaken as part of a project by students, postdocs and faculty at the Immunology Institute of the Icahn School of Medicine, Mount Sinai.

On 2020-03-26 00:17:43, user Peter Dunn wrote:

Interesting study, but the abstract needs some editing. There are words missing. eg " <br /> SAK models more logistically feasible" should be SAK models WERE more logistically feasible.

On 2020-03-25 23:20:42, user Gill wrote:

Could you comment on naming At4g15750 PMEI13. According to TAIR and other papers eg www.plantcell.org/cgi/doi/1..., AT5G62360 is PMEI13?

On 2020-03-25 22:48:40, user Dan Seipel wrote:

Would you expect TUDCA to have a similar effect?

On 2020-03-25 22:31:23, user Sinai Immunol Review Project wrote:

Title

LY6E impairs coronavirus fusion and confers immune control of viral disease

Keywords

interferon-stimulated genes, antiviral interferons, human coronaviruses (CoV), murine hepatitis virus (MHV)

Key findings

Screening a cDNA library of >350 human interferon-stimulated genes for antiviral activity against endemic human coronavirus HCoV-229E (associated with the common cold), Pfaender S & Mar K et al. identify lymphocyte antigen 6 complex, locus E (Ly6E) as an inhibitor of cellular infection of Huh7 cells, a human hepatoma cell line susceptible to HCoV-229E and other coronaviruses. In a series of consecutive in vitro experiments including both stable Ly6E overexpression and CRISPR-Cas9-mediated knockout the authors further demonstrate that Ly6E reduces cellular infection by various other coronaviruses including human SARS-CoV and SARS-CoV-2 as well as murine CoV mouse hepatitis virus (MHV). Their experiments suggest that this effect is dependent on Ly6E inhibition of CoV strain-specific spike protein-mediated membrane fusion required for viral cell entry.

To address the function of Ly6E in vivo, hematopoietic stem cell-specific Ly6E knock-out mice were generated by breeding Ly6Efl/fl mice (referred to as functional wild-type mice) with transgenic Vav-iCre mice (offspring referred to as Ly6E HSC ko mice); wild-type and Ly6E HSC ko mice of both sexes were infected intraperitoneally with varying doses of the natural murine coronavirus MHV, a pathogen that depending on route of infection can cause a wide range of diseases in mice including hepatitis, enteritis and encephalomyelitis. Briefly, compared to wild-type controls, mice lacking hematopoietic cell-expressed Ly6E were found to present with increased mortality, a more severe disease phenotype as based on serum ALT levels (prognostic of liver damage), liver histopathology, and viral titers in the spleen. Moreover, bulk RNAseq analysis of infected liver and spleen tissues indicated changes in gene expression pathways related to tissue damage and antiviral immune responses as well as a reduction of genes associated with type I IFN response and inflammation. Finally, the authors report substantial differences in the numbers of hepatic and splenic APC subsets between wild-type and knockout mice following MHV infection and show that Ly6E-deficient B cells and to a lesser extent also DCs are particularly susceptible to MHV infection in vitro.

Potential limitations

Experiments and data in this study are presented in an overall logical and coherent fashion; however, some observations and the conclusions drawn are problematic and should be further addressed & discussed by the authors. Methodological & formal limitations include relatively low replicate numbers as well as missing technical replicates for some in vitro experiments (cf. Fig. legend 1; Fig. legend 2e); the omission of “outliers” in Fig. legend 2 without an apparent rationale as to why this approach was chosen; the lack of detection of actual Ly6E protein levels in Ly6E HSC ko or wild-type mice; and most importantly, missing information on RNAseq data collection & analysis in the method section and throughout the paper. A more relevant concern though is that the interpretation of the experimental data presented and the language used tend to overrate and at times overgeneralize findings: for example, while the authors demonstrate statistically significant, Ly6E-mediated reduction of coronavirus titers in stable cells lines in vitro, it remains unclear whether a viral titer reduction by one log decade would be of actual biological relevance in face of high viral titers in vivo. After high-dose intraperitoneal MHV infection in vivo, early viral titers in Ly6E HSC knockout vs. wt mice showed an elevation in the spleen (~1.5 log decades) but not liver of the ko mice (other tissue not evaluated), and while ko mice presented with only modestly increased liver pathology, both male and female ko mice exhibited significantly higher mortality. Thus, the manuscript title statement that “Ly6E … confers immune control of viral disease” is directly supported by only limited in vivo data, and gain-of-function experiments (eg. Ly6E overexpression) were not performed. Of additional note here, tissue tropism and virulence differ greatly among various MHV strains and isolates whereas dose, route of infection, age, genetic background and sex of the mice used may additionally affect disease outcome and phenotype (cf. Taguchi F & Hirai-Yuki A, https://doi.org/10.3389/fmi... Kanolkhar A et al, https://jvi.asm.org/content/ 83/18/9258). Observations attributed to hematopoietic stem cell-specific Ly6E deletion could therefore be influenced by the different genetic backgrounds of floxed and cre mice used, and although it appears that littermates wt and ko littermates were used in the experiments, the potentially decisive impact of strain differences should at least have been discussed. Along these lines, it should also be taken into account that the majority of human coronaviruses cause respiratory symptoms, which follow a different clinical course engaging other primary cellular mediators than the hepatotropic murine MHV disease studied here. It therefore remains highly speculative how the findings reported in this study will translate to human disease and it would therefore be important to test other routes of MHV infection and doses that have been described to produce a more comparable phenotype to human coronavirus disease (cf. Kanolkhar A et al, https://jvi.asm.org/content/ 83/18/9258). Another important shortcoming of this study is the lack of any information on functional deficits or changes in Ly6E-deficient immune cells and how this might relate to the phenotype observed. Overall, the in vitro experiments are more convincing than the in vivo studies which appear somewhat limited.

Overall relevance for the field

Despite some shortcomings, the experiments performed in this study suggest a novel and somewhat unexpected role of Ly6E in the protection against coronaviruses across species. These findings are of relevance and should be further explored in ongoing research on potential coronavirus therapies. Yet an important caveat pertains to the authors’ suggestion that “therapeutic mimicking of Ly6E action” may constitute a first line of defense against novel coronaviruses since their own prior work demonstrated that Ly6E can enhance rather than curtail infection with influenza A and other viruses.

Reviewed as part of a project by students, postdoctoral fellows and faculty at the Immunology Institute of the Icahn School of Medicine at Mount Sinai

On 2020-03-25 21:10:56, user Sinai Immunol Review Project wrote:

Summary of Findings: <br /> - Used human liver ductal organoids to determine ACE2+ cholangiocytes in healthy liver (2.92% of all cells) are infectable and support SARS-CoV-2 viral replication. <br /> - Plaque-purified SARS-CoV-2 viral infection disrupted organoid barrier and bile transporting functions of cholangiocytes through dysregulation of genes involved in tight junction formation (CLDN1) and bile acid transportation (ASBT and CFTR).

Limitations: <br /> - Unclear if liver damage observed in patients due to direct cholangiocyte infection or due to secondary immune/cytokine effects.<br /> - This study argues for direct damage as it lacks immune contexture; but further studies needed with autopsy samples or organoid-immunecell co-culture to conclude strongly. <br /> - Would be important to measure cholangiocyte-intrinsic anti-viral response and alarmins secreted upon infection, and furthermore study tropism of various immune cells to conditioned media from organoids infected with SARS-CoV-2. <br /> - Does not address how cirrhotic- or alcohol/smoking/obesity-associated liver organoids respond to SARS-CoV-2 infectivity and replication, worth pursuing to experimentally address clinical data indicating co-morbidities.

Importance/Relevance: <br /> - Useful model to rapidly study drug activity against SARS-CoV-2 infection in liver, while monitoring baseline liver damage.

• Liver abnormality observed in >50% of CoVID-19 patients; the results from this study could explain the bile acid accumulation and consequent liver damage observed.

Review by Samarth Hegde as part of a project by students, postdocs and faculty at the <br /> Immunology Institute of the Icahn school of medicine, Mount Sinai.

On 2020-03-25 17:02:59, user Sinai Immunol Review Project wrote:

Summary: The authors use 2 neural network algorithms, NetMHCpan4 and MARIA, to identify regions within the COVID-19 genome that are presentable by HLA. They identify 405 viral epitopes that are presentable on MHC-I and MHC-II and validate using known epitopes from SARS-CoV. To determine whether immune surveillance drives viral mutations to evade MHC presentation, the authors analyzed 68 viral genomes from 4 continents. They identified 93 point mutations that occurred preferentially in regions predicted to be presented by MHC-I (p=0.02) suggesting viral evolution to evade CD8 T-cell mediated killing. 2 nonsense mutations were also identified that resulted in loss of presentation of an associated antigen (FGDSVEEVL) predicted to be good antigen for presentation across multiple HLA alleles. <br /> To identify potential sites of neutralizing antibody binding, the authors used homology modeling to the SARS-CoV’s spike protein (S protein) to determine the putative structure of the CoV2 spike protein. They used Discotope2 to identify antibody binding sites on the protein surface in both the down and up conformations of the S protein. The authors validate this approach by first identifying antibody binding site in SARS-CoV S protein. In both the down and up conformation of the CoV2 S protein, the authors identified a potential antibody binding site on the S protein receptor binding domain (RBD) of the ACE2 receptor (residues 440-460, 494-506). While RBDs in both SARS-CoV and CoV2 spike proteins may be important for antibody binding, the authors note that SARS-CoV has larger attack surfaces than CoV2. These results were later validated on published crystal structures of the CoV2 S protein RBD and human ACE2. Furthermore, analysis of 68 viral genomes did not identify any mutations in this potential antibody binding site in CoV2. <br /> Finally, the authors compile a list of potential peptide vaccine candidates across the viral genome that can be presented by multiple HLA alleles. Several of the peptides showed homology to SARS-CoV T-cell and B-cell epitopes.

Limitations: While the authors used computational methods of validation, primarily through multiple comparisons to published SARS-CoV structures and epitopes, future work should include experimental validation of putative T-cell and B-cell epitopes.

Importance: The authors identified potential T-cell and B-cell epitopes that may be good candidates for peptide based vaccines against CoV2. They also made interesting observations in comparing SARS-CoV and CoV2 potential antibody binding sites, noting that SARS-CoV had larger attack surfaces for potential neutralizing antibody binding. One of the highlights of this paper was the authors’ mutation analysis of 68 viral genomes from 4 continents. This analysis not only validated their computational method for identifying T-cell epitopes, but showed that immune surveillance likely drives viral mutation in MHC-I binding peptides. The smaller attack surface may point to potential mechanisms of immune evasion by CoV2. However, absence of mutations in the RBD of CoV2 and the small number of mutations in peptides presentable to T cells suggests that vaccines against multiple epitopes could still elicit robust immunity against CoV2.

On 2020-03-25 15:30:06, user Sinai Immunol Review Project wrote:

Rhesus macaques were immunized intramuscularly twice (week 0 and week 4) with SV8000 carrying the information to express a S1-orf8 fusion protein and the N protein from the BJ01 strain of SARS-CoV-1. By week 8, immunized animals had signs of immunological protection (IgG and neutralization titers) against SARS-CoV-1 and were protected against challenge with the PUMC-1 strain, with fewer detectable symptoms of respiratory distress, lower viral load, shorter periods of viral persistence, and less pathology in the lungs compared to non-immunized animals.

The authors should write clearer descriptions of the methods used in this article. They do not describe how the IgG titers or neutralization titers were determined. There are some issues with the presentation of data, for example, in Figure 1a, y-axis should not be Vmax; forming cells and 1d would benefit from showing error bars. Furthermore, although I inferred that the animals were challenged at week 8, the authors did not explicitly detail when the animals were challenged. The authors should explain the design of their vaccine, including the choice of antigens and vector. The authors also do not include a description of the ethical use of animals in their study.

The authors describe a vaccine for SARS-CoV-1 that could benefit from a discussion of possible implications for the current SARS-CoV-2 pandemic. Could a similar vaccine be designed to protect against SARS-CoV-2 and would the concerns regarding emerging viral mutations that the authors describe as a limitation for SARS-CoV-1 also be true in the context of SARS-CoV-2?

On 2020-03-25 15:04:57, user Matteo Paccagnella wrote:

Hi I'd like to know if it'll be possible use a DESI ionization technique to avoid cromatography separation, maybe it could be faster processing larger number of samples? Maybe spectra would be too complicated...

On 2020-03-22 15:43:57, user Biggest_of_Bears wrote:

are there mass spec parameters for the conditions required (Columns, temps, mobile phase gradients, ion sources, and transitions?). I would like to see if this is possible on a 6460 triple quad with the components I have available in my lab

On 2020-03-25 13:48:59, user Melanie Haffner-Luntzer wrote:

This fits perfectly to our finding that OVX delays chondrocyte-to-osteoblast transformation during fracture healing and that there is less beta-catenin in hypertrophic chondrocytes in the fracture callus of OVX mice! https://dmm.biologists.org/...

On 2020-03-25 09:39:19, user x wrote:

"Sp/Nsp cocktail vaccine" containing a structural protein(s) (Sp) and a non-structural protein(s) (Nsp) would stimulate effective complementary immune responses!!!!

On 2020-03-25 00:25:15, user Sinai Immunol Review Project wrote:

Summary of Findings:<br /> - To investigate the possible cause of kidney damage in 2019-nCoV patients, authors used published kidney and bladder cell atlas data (GSE131685, GSE108097; 3 healthy donors each) as well as an unpublished kidney single-cell RNA-Seq data (in-house from 2 transplant donors) to evaluate ACE2 gene expressions in all cell types of healthy kidneys and bladders. <br /> - They find enriched expression of ACE2 transcript in all subtypes of proximal tubule cells of kidney, with 5%-15% of both straight and convoluted proximal tubule cells expressing ACE2. <br /> - They find detectable levels of ACE2 in bladder epithelial cells, noting expression from around 1.5% of cells in the outer layer umbrella cells of the bladder epithelium and decreasing in the basal cells. <br /> - Importantly endothelial or immune cells in kidney/bladder do not express ACE2.

Limitations:<br /> - This study primarily characterizes ACE2 expression (amongst other genes) from a small healthy-donor dataset, and will benefit from supporting data in (expired) patient samples to show functional viral damage. ACE2 transcript does not necessarily translate to viral permissiveness in kidney/bladder epithelia or cytokine release. <br /> - This study focuses on only healthy tissue; it will be useful to analyze kidney/bladder epithelial ACE2 expression under inflammatory conditions or in patients with underlying kidney conditions. <br /> - Given what is known about protease TMPRSS2 expression during SARS-CoV-2 infection, ACE2+TMPRSS2+ double-positive cell identification would be useful in these datasets.

Importance/Relevance:

• ACE2 protein is spatially restricted to brush border of proximal tubules and in bladder umbrella cells [1], such cells in direct contact with viral particles are likely to be highly sensitive to viral-induced damage.
• SARS-CoV and MERS-CoV have been shown to be detected in urine of patients and associate with higher mortality [2,3], thus worth understanding kidney damage and resultant immune response in SARS-CoV-2 as well.
• This study argues for a potential mode of viral infectivity and resultant inflammatory responses in these tissue in addition to reported infectivity in the lung and digestive system, which is supported by clinical data showing acute and early kidney complications in 2019-nCoV patients (https://doi.org/10.1101/202...
• Clinically, thus very important to track urinary CoVID-19 shedding as well as study acute kidney injury-related co-morbidities.

References:

1. Hamming I, Timens W, Bulthuis ML, et al. The Journal of pathology 2004;203(2):631-7. doi:10.1002/path.1570
2. Chu KH, Tsang WK, Tang CS, et al. Kidney international 2005;67(2):698-705. <br /> doi:10.1111/j.1523-1755.2005.67130
3. Arabi YM, Arifi AA, Balkhy HH, et al. Annals of internal medicine 2014;160(6):389-97. <br /> doi: 10.7326/m13-2486

Review by Samarth Hegde as part of a project by students, postdocs and faculty at the <br /> Immunology Institute of the Icahn school of medicine, Mount Sinai.

On 2020-03-24 18:25:53, user Katrina wrote:

Review of “Photosynthetic protein classification using genome neighborhood-based machine learning feature” by Sangphukieo et. Al. provided by the discussions during a computational biology journal club at University of Tennessee at Knoxville

Summary: <br /> The authors find that their genome neighborhood-based model can identify known photosynthetic proteins with 94% accuracy, 0.892 F1 minor which is higher accuracy than prior sequence-based models and blastp methods. The novelty of this work is gene neighborhood network feature extractions. These features are generated by determining gene neighbors by summing the total branch length of the tree for each shared gene content between the 154 genomes normalized by a quartile cutoff. They compare their method to several other classification of photosynthetic gene software methods, feature reduction methods and different model classifiers.

The author’s claim that identifying photosynthesis related genes is hard because photosynthetic components are temporarily present in plants, experimental identification costs time and money. The motivation is to improve photosynthetic efficiency since the current photosynthesis efficiency is at 6%. Previous computational approaches rely on sequence similarity and can falsely label genes as photosynthesis related. Machine learning provides a non-homology way to identify photosynthesis genes. No previous computational methods for predicting protein function incorporating gene cluster information. The dataset was tested on photosynthetic and non-photosynthetic proteins from UniprotKB. They propose this work includes genomic context and sequence similarity, two criteria that may be useful to identifying function for photosynthetic genes.

Major comments:

What is the exact purpose for this work? Is it to create an essential cyanobacteria genome, or understanding the genes involved in photosynthesis in cyanobacteria? What is the fundamental knowledge that we are learning from identifying photosynthetic genes vs. non-photosynthetic genes?

How were proteins from UniprotKB annotated to be photosynthetic vs. non-photosynthetic? Is this based on GO ontology? Is GO ontology best for verifying photosynthetic genes or is there a better metric/dataset to use? How does filtering types of relationships based GO ontology impact accuracy of methods? What about experimental verified GO ontology relationships included? Are experimental only verified relationships have high accuracy of identifying photosynthetic genes?

How do the results work if used only experimentally verified photosynthetic proteins vs. GO annotated? Were the labels too generic? And how can go ontology be incorrect and maybe better label better?

It is not clear what data the tree was built on to determine phylo score? The proteomes of all 154 genomes , the shared genes only between 154 proteomes, whole genome sequence?

The thresholds for the e-value are very high to get high accuracy for predicting novel photosynthetic proteins (greater than 1)?

How were 154 genomes selected from NCBI? How are representatives from each 7 phyla determined? What NCBI database? RefSeq, Genbank, SRA? There are over 10,000 cyanobacteria genomes in Genbank.

Minor comments:

On line 38, the motivation is to improve photosynthetic efficiency since the current photosynthesis efficiency is at 6%. What does 6% efficiency mean for photosynthesis?

For table 2, were there duplicate features included in the model when combining all features for all e-value cutoffs (line 248) or were duplicate features removed?

Based on S3, the minimization of features figure, shows a trend of higher accuracy, F1 minor and MCC as number of features increase not as number of features decrease. Please check lines 241-243 again. The data does not support feature reduction and show it is helping model prediction.

On line 160, the split of data 90% training model and 10% for testing. Why this cutoff? Usually the cutoff is 60/40? Is the model overfit for this data?

On line 138, the normalization method is applied for phyloscore. Why not divide the phylo score by average phylo score to normalize? Please explain rationale for quartile method and how this could impact results by setting different cutoffs and what is quartile cutoff for level 2?

In table 3 on line 533, the recall and precision would be good for photomod for identifying known photosynthetic genes. Why weren't those metrics included?

On line 335, there are all sequence based methods. Can you confirm why you think your method is not sequence based ? <br /> On line 312, put S9 where validated the table. <br /> On line 116- 117, what is the rationale for this ?

Why does neighboring genes have to be conserved in at least 3 genomes? Rationale for this cutoff and how this was determined?

On line 101, what was the rationale for blastp cutoffs of 1E-10, 1E-50 and 1E-100?

On line 149, you state there are exactly 6,430 photosynthetic and 6,430 non-photosynthetic proteins. Is this true statement for the equal number of photosynthetic proteins and non-photosynthetic proteins?

The f1 minor is only for classifying photosynthetic genes. Do you think it is important to classify non-photosynthetic genes and have F1 including this too into metric? Justify use of the F1 minor metric and that it may over stress noise and why not use F1? What is the F1?

For readers, can you include supplemental at the end of the main file and include the figures incorporated into the main text? This will help readers comment and follow better.

The current colors in the figure are not legible if paper is printed out in black and white.

The current figures are rasterized images. Can you please make it vectorized for scalable images? Ex. Figure 3. Also you can include full high resolution at the end of file and low resolution images in the paper text itself.

Thank you for including line numbers

On 2020-03-24 15:57:40, user Brice Curtin wrote:

Can you comment on why the lactam appears as a hemiaminal in the crystal structure? This is shown in Fig. 2 where the geometry is not that of a lactam, and the 3D structure wouldn't match the covalently-modified structure in Extended Data Figure 1c. This change in structure was not discussed in the text.

On 2020-03-24 13:33:37, user Anton Nekrutenko wrote:

"The resulting data was analyzed by Paragon Genomics’ Bioinformatics team with proprietary pipelines and algorithms." - this is hardly a way to address pandemic data analysis needs

On 2020-03-24 12:25:01, user Sinai Immunol Review Project wrote:

Vaccination of mice with a single dose of a 9-amino-acid peptide NP44-52 located in a conserved region of ebolavirus (EBOV) nucleocapsid protein (NP) confers CD8+ T-cell-mediated immunity against mouse adapted EBOV (maEBOV). Bioinformatic analyses predict multiple conserved CD8+ T cell epitopes in the SARS-CoV-2 NP, suggesting that a similar approach may be feasible for vaccine design against SARS-CoV-2.

The authors focus on a site within a 20-peptide region of EBOV NP which was commonly targeted by CD8+ T cells in a group of EBOV survivors carrying the HLA-A*30:01:01 allele. To justify the testing of specific vaccine epitopes in a mouse challenge setting, the authors cite known examples of human pathogen-derived peptide antigens that are also recognized by C57BL/6 mice, as well as existing data surrounding known mouse immunogenicity of peptides related to this EBOV NP region. Testing 3 distinct 9mer peptides over an 11 amino-acid window and comparing to vaccination with the 11mer with a T-cell reactivity readout demonstrated that optimizing peptide length and position for immunogenicity may be crucial, likely due to suboptimal peptide processing and MHC-class-I loading.

Vaccines for maEBOV challenge studies were constructed by packaging NP44-52 in d,l poly(lactic-co-glycolic) acid microspheres. CpG was also packaged within the microspheres, while Monophosphoryl Lipid A (a TLR4 ligand) was added to the injectate solution. A second peptide consisting of a predicted MHC-II epitope from the EBOV VG19 protein was added using a separate population of microspheres, and the formulation was injected by intraperitoneal administration. The vaccine was protective against a range of maEBOV doses up to at least 10,000 PFU. Survival was anticorrelated with levels of IL6, MCP-1 (CCL2), IL9, and GM-CSF, which recapitulated trends seen in human EBOV infection.

While HLA-A*30:01:01 is only present in a minority of humans, the authors state that MHC binding algorithms predict NP44-52 to be a strong binder of a set of more common HLA-A*02 alleles. The authors predict that a peptide vaccine based on the proposed formulation could elicit responses in up to 50% of people in Sudan or 30% of people in North America. <br /> SARS-CoV-2 NP, meanwhile, has conserved regions which may provide peptide-vaccine candidates. Scanning the SARS-CoV-2 NP sequence for HLA-binding 9mers identified 53 peptides with predicted binding affinity < 500nM, including peptides that are predicted to bind to HLA-class-I alleles of 97% of humans, 7 of which have previously been tested in-vitro.

The results support previously appreciated correlations between certain cytokines and disease severity, specifically IL6 which relates to multiple trial therapies. Prediction of HLA-class-I binding of SARS-CoV-2 NP peptides suggests the plausibility of a peptide vaccine targeting conserved regions of SARS-CoV-2 NP although further validation in previously infected patient samples will be essential.

Review by Andrew M. Leader as part of a project by students, postdocs and faculty at the Immunology Institute of the Icahn school of medicine, Mount Sinai.

On 2020-03-24 12:20:52, user Surisetty V V Arun Kumar wrote:

It seems Appendix-1 referrred in the text is missing in the pdf

On 2020-03-24 00:43:32, user Nethaneel Mongonia wrote:

largely problematic with this study, is that it doesn’t account for societal gender roles, in the development of muscle through labor, and sports. Individuals assigned male at birth, are more likely to be encouraged to participate in strength training, manual labor, and competitive sports. This provides a huge advantage in starting position, regardless of sex hormones. A women that has been power lifting for years, will out perform a male with no history of physical labor, or athletic sports history. So it would reason, that taking a male and a female off the street, would be an awful barometer. A better designed study would include individuals sedentary assigned males transitioning to female, and athletic assigned female transitioning to male. And so forth. This would give us a larger picture of how social gender roles effect muscle development, and attempt to adjust for said environmental factors. This study is just lazy.

On 2020-03-23 16:04:02, user Mohamed F. Sallam wrote:

Kindly contact the corresponding author for updated version of the manuscript. Lots of improvement have been done since submission on Biorxiv.

On 2020-03-23 14:10:14, user Cristobal Uauy wrote:

This is a really great resource. Unfortunately I was unable to retrieve the data through the provided BioProject number (PRJNA607017). Would it be possible for the authors to make this data accessible to the research and breeding community?

On 2020-03-23 12:47:00, user Ben Berman wrote:

is sgRNA the typical way to refer to subgenomic RNA? For people new to the field , it is confusing with CRISPR sgRNA (single guide RNA). Thanks

On 2020-03-23 03:32:41, user ppgardne wrote:

Looks like a very interesting preprint. Is there a supplementary material? In particular, where is Table S5 available?

On 2020-03-22 17:28:12, user Lorenzo Calviello wrote:

Hi, great paper, thanks!

Do you see longer poly(A) tails (e.g. Figure 4C) on host mRNAs during infection? Asking because of this study, where transfecting the NSP1 protein from SARS-CoV induces gene expression downregulation via hyperadenylation and nuclear retention of mRNAs: https://mcb.asm.org/content... (Figure 7)

Thanks again!<br /> Lorenzo

On 2020-03-23 12:31:56, user UAB Journal Club wrote:

The Bacterial Pathogenesis and Physiology Journal Club<br /> The University of Alabama at Birmingham<br /> Spring 2020

Review of “Filamentous bacteriophage delay healing of Pseudomonas-infected wounds”<br /> Bach et al., BioRXiV

Summary<br /> In this manuscript, the authors follow up on their earlier work implicating the filamentous Pf phage in Pseudomonas aeruginosa virulence, with a study focused on the effects of that phage on wound healing in both mice and humans. They establish clear correlations between the presence of Pf phage and delayed healing, identifying preliminary mechanistic explanation for this effect (impaired keratinocyte migration and alteration of monocyte/macrophage wound response).

We have found this paper to be well-written and the results and methods to be clearly described. While we would like to offer some suggestions to the authors and pose questions that have come forth, overall, we find the authors’ conclusions convincing.

Minor Comments<br /> 1) The paragraph in the Results section entitled “Pf phage enhance in vitro Pa biofilm formation” describes results that seem to be a replication of previously published findings (see Rice et al. 2008 ISME J 3:271–282). Please at least acknowledge this.

2) The authors to not indicate how much Pa each mouse is inoculated with. Include in figure legends or methods.

3) The authors should include scale bars on all confocal images

4) The authors did not define the term “epithelial gap”. The manuscript would benefit from this inclusion.

Major Comments<br /> 1) The authors establish that soluble factors produced by Pf-exposed macrophages impact keratinocyte migration (Fig. 3), but are not able to replicate these effects with purified components. This is reported as “data not shown”, but we think this is a very important result. We would like to see these data, and even if the authors are not able to establish which of the soluble factors are important in this effect (experiments which would greatly increase the impact of this paper, but are, admittedly, challenging), would like to have more discussion of what those factor(s) might be and how the authors intend to solve this problem.

2) Data regarding replicates, statistics and methods were improperly analyzed, nor discussed, or not reported in certain sections:

a. For Fig. 3 there are not enough independent replicates for the experiments in panels C and F to do the reported statistical tests.

b. No information is given for the number of replicates for the experiments shown in Fig. 4. Please ensure that at least 3 independent replicates of each experiment are performed.

c. Broad consideration of appropriate statistical analysis should be assessed. Statistical analyses for repeated measures data should be considered whenever data are in time-course, one-way ANOVA is not appropriate for these data. Additionally, the citation of which post-hoc analysis was used for all ANOVA should be included in figure legends. Replicate (N, population number) should be included in all reporting of statistics within figure legends (for example, replicate number and how data are combined is not explained in Figure 4B). ANOVA should be used to assess three-group comparisons, rather than paired or unpaired student t-test (e.g. Figures 2 and 3).

d. Methods for the phage uptake experiments are lacking detail (e.g. Flow protocol, stains used for Live/dead, etc.)

e. In the results section title “Pf phage alter macrophage cytokine production profiles,” it is unclear that the sample being tested is the supernatant of the BMDMs treated with Pf or LPS. While this is explained in the figure legend, the manuscript would benefit from clarifying the experimental procedure in the text of the results section.

3) The authors discuss briefly the need to know whether Pf phage are found in Pseudomonas isolated from other infection sites, but we would argue this is an important piece of information for interpreting the human wound dataset.

a. Are Pf phage actually more abundant in wound isolates than in other infectious strains? Are Pf phage common in environmental isolates of P. aeruginosa? While an exhaustive survey is probably beyond the scope of this paper, a good baseline could be obtained using the author’s established qPCR method on a variety of P. aeruginosa strains available from culture collections.

b. Are Pf phage present in mucoid strains of P. aeruginosa associated with other pathologies? (a discussion point which my broaden the application of these results

Figure-specific comments:

1) Figure 1: 1B, 1G, and 1H would benefit from inclusion of objective scoring in addition to the representative images as this is a major foundation of the study, allowing for statistical comparisons.1D and 1E both include day 13 timepoint but with different N’s. This is concerning and should be mentioned in the legend if they are from separate experiments.

2) Figure 2: it would benefit the conclusions to see a time course of uninfected wounds day 1-13 compared to the day 1-13 infection model. Additionally, there is no representative image for the PF4 treated wound, an important piece of data considering the normalization to wound area on Day 1. We suggest adding the data which show no alteration in inflammatory cell counts in Pf phage treated groups is important enough to include in the main figure 2 rather than in supplemental figures. This is an important foundation to build that point Pf4 phage is impairing re-epithelialization of the wound rather than altering inflammatory responses.

3) Figure 3: 3C is lacking a positive control. It is reported that there were only two experiments assessed which would not be properly assessed by ANOVA. Clarification of replicates and sample size or how statistics were used in either figure legend or broadly in the methods would clarify. It is unclear why representative data are shown in Fig. 3B. Panels 3C and 3F represent the same output data measures and thus the axes should be the same as to not distort the data. Inclusion of a positive control to 3C would resolve this.

4) Figure 4: The authors do not specify the selection for further analysis of certain cytokines/chemokines from the larger dataset from Figure 4B. For examples, VCAM1 was selected while TNFa was not, despite a clear reduction in TNFa levels in the Pf-treated group. A brief mention of reasoning for selection would improve understanding of results. It is mentioned that many of these factors were tested but the data was not shown. If these data are negative, it would be valuable to include these in a table form with the limit of detection listed for each within the supplemental (e.g. TNFa <7pg/mL). There is a panel for TGFa (which was not discussed in results) but not the TNFa which was. Verify data shown and clarify in discussion.

5) Figure 5: Panels do not match the reference/legend (e.g. i.e. H in figure, I in caption)

On 2020-03-23 08:43:24, user Masa Tsuchiya wrote:

We have updated genomic mechanism for cell-fate determination based on time-series whole gene expression data analysis.

In this revision, from embryo to cancer development, the following five points are clearly provided for the development of a biological regulation transition theory demonstrating the cell-fate change:

1. There exists a center of cell-fate in the genome, which corresponds to a critical point (CP) and determines cell-fate change.

2. The CP is the genome-attractor. The switching singular behaviors at the CP transforms the genome into a super-critical state (i.e., super-critical genome).

3. In the super-critical genome, a specific stochastic perturbation can spread over the entire system through the ‘genome engine’ - an autonomous critical-control genomic system. Whereas, in the sub-critical state (no transitional change in the CP), the perturbation remains at a local level.

4. The super-critical genome induces a ‘global expression avalanche’ over the whole expression to guide the cell-fate change.

5. Cell-fate change occurs after the genome passes over a stable point (non-equilibrium fixed point) of the thermodynamically open system. The coherent perturbation on the genome-engine derives cell-fate change including reprogramming in embryo development.

On 2020-03-22 13:47:52, user Vahe Demirjian wrote:

The Oculudentavis paper raises questions about when birds started achieving miniature size because the recently described basal avialan Fukuipteryx (Imai et al. 2019), also recovered as more basal than Jeholornis, is bigger than Oculudentavis, and so is the taxon Yandangornis (Cai & Zhao 1999).

Cai, Z. & Zhao, L. (1999). "A long tailed bird from the Late Cretaceous of Zhejiang". Science in China Series D-Earth Sciences Vol. 42, No. 4, 1999 pp.434-441. ISSN 1006-9313

Imai, T., Azuma, Y., Kawabe, S., Shibata, M., Miyata, K., Wang, M., & Zhou, Z., 2019. An unusual bird (Theropoda, Avialae) from the Early Cretaceous of Japan suggests complex evolutionary history of basal birds. Communications Biology, 2(1). doi: 10.1038/s42003-019-0639-4

On 2020-03-20 07:08:08, user Martin Bäker wrote:

Question concerning lines 57/58: Is "antorbital fenestra" a typo for "orbit"?

On 2020-03-22 10:23:52, user William P Hall wrote:

A most interesting paper. I wish that I had had access to this kind of technology when I was studying the chromosomal variation in Sceloporus.

Something you might consider is that based on chaisma frequency in most undulatus group species I karyotyped is that this was typically very low, often only one per chromosome pair as documented from surviving chromosome cards (for transcribed records, see https://www.dropbox.com/s/e.... From 50 year old memory in my 80 year old brain, in the 6 pairs of macrochromosomes, the single chiasma was normally located at or close to the centromere, suggesting that little recombination took place in the chromosome arms. I don't recall any particular exceptions for the 7th pair that was relatively larger than the typically small microchromosomes of most other species with higher chromosome numbers. In a few other species pair 7 was heteromorphic or involved in an X1, X2, Y sex chromosome trisomy.

Sceloporus clearly has a lot more to reveal where the evolution and function of genetic systems is concerned.

On 2020-03-22 10:20:55, user Yu Lan wrote:

The HECs we identified showed unambiguous endothelial characteristics molecularly, functionally, and anatomically. Furthermore, the hematopoietic feature was largely ruled out as the HECs hardly generated CFU-Cs by directly cultured in methylcellulose culture medium. We have been aware of the potential low CD41 expression in HECs at both protein (by FACS antibody labelling) and transcriptional level. However, as Itga2b (encoding CD41) was ubiquitously low expressed in many embryonic ECs in addition to HECs at least transcriptionally, we did not consider the low expression of CD41 as a criterion of non-ECs. For the above reasons and from the view of vascular EC evolution, we prefer to name them as HECs but not pro-HSCs. Of note, the HECs we identified showed an obviously proliferative status whereas pro-HSCs are slowly cycling as reported. We are very pleased to directly and comprehensively compare the pro-HSCs and our HECs in the near future.