Dear Prof. Palsson,

Thank you very much for submitting your manuscript "Matrix factorization recovers consistent regulatory signals from disparate datasets" for consideration at PLOS Computational Biology.

Our apologies for the delay in decision. We have had difficulties in securing reviews on time - likely due to the current pandemic situation that has placed much burden on the academic community. We thank you for your patience and understanding.

Your manuscript was reviewed by members of the editorial board and two independent reviewers. In light of the reviews (see my own comments next and the reviewers' comments below this email), we would like to invite the resubmission of a significantly-revised version that takes into account the reviewers' comments.

We cannot make any decision about publication until we have seen the revised manuscript and your response to the reviewers' comments. Your revised manuscript is also likely to be sent to reviewers for further evaluation.

When you are ready to resubmit, please upload the following:

[1] A letter containing a detailed list of your responses to the review comments and a description of the changes you have made in the manuscript. Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

[2] Two versions of the revised manuscript: one with either highlights or tracked changes denoting where the text has been changed; the other a clean version (uploaded as the manuscript file).

Important additional instructions are given below your reviewer comments.

Please prepare and submit your revised manuscript within 60 days. If you anticipate any delay, please let us know the expected resubmission date by replying to this email. Please note that revised manuscripts received after the 60-day due date may require evaluation and peer review similar to newly submitted manuscripts.

Thank you again for your submission. We hope that our editorial process has been constructive so far, and we welcome your feedback at any time. Please don't hesitate to contact us if you have any questions or comments.

Sincerely,

Kiran Raosaheb Patil, Ph.D.

Associate Editor

PLOS Computational Biology

Weixiong Zhang

Deputy Editor

PLOS Computational Biology

***********************

Editor comments:

Major comments:

1. There is no analysis/discussion regarding the differences that can (and most likely will) arise from grouping different strains together. Strains resulting from ALE can harbour substantial differences in regulatory program, and this effect can be much greater in natural variants (even the genome size can differ by >10%). I recommend to discuss this, whenever possible with supporting additional analysis.

2. Given that future application for the method is going to be either much larger datasets or other species (as authors point out), I think the manuscript would much benefit by expanding the scope in this direction. This could be done, e.g., by analysing datasets for additional species of health/industrial importance, and/or by including cross-species comparison, and/or by increasing the number of datasets. This would also help addressing the questions of novelty raised by the other reviewer.

Minor comments:

1. Fig 2: X-axis labels unclear. R would be better indicated inside the plot and not as a x-axis label.

2. Table 1: Please consider removing the column ‘Research Group’. This is redundant with the next column (reference) and also unnecessarily gives an impression of subjectivity.

Reviewer's Responses to Questions

Comments to the Authors:

Please note here if the review is uploaded as an attachment.

Reviewer #1: The paper by Sastry et al presents some results based on the use of ICA to recover regulatory signals from different datasets. The topic is a relevant one, and overall the study presented is, in my opinion, technically sound and presents some technical details useful for the community. Still, I believe that the results of the paper do not represent a major breakthrough to warrant its publication in PLOS CB.

Indeed, the main contribution of the paper over previous work is, as stated in the last paragraph of the Introduction, to show that the method provided in previous work based on ICA is capable of retrieving consistent regulatory modules from different transcriptomics and proteomics datasets for E. coli. Although some results point to consistent modules, I believe that the work shown here is still preliminary, showing somewhat marginal contributions over previous work, and some of the claims need more support. Some reasons for this statement are listed below:

- the analysis of the consistence of the iModulons obtained in based on Pearson correlation of the coefficients of the ICA (a metric of linear correlation while ICA can capture nonlinear relationships; maybe mutual information would be an option);

- a match of different modules in different datasets is obtained based on reciprocal best hit; no analysis is provided on the number of shared genes. Also, RBH is done pairwise and thus the 45% of linked iModulons is an interesting result, but not of a level that justifies some of the produced claims. The number of modules consistent in all 5 datasets (or even 3-4) seems to be low (even the example presented of CysB only holds for 4 datasets), although this is not discussed.

- the RBH by itself does not represent a high consistency between the connected modules, only that they are the best candidates to be connected.

- it is not obvious why almost half of the modules retrieved from microarrays datasets cannot be characterised.

- the results shown for the proteomics datasets seem to be very inconclusive and could be better discussed; they do not seem to justify the initial claim (in the abstract: "echoes of this structure remain in the proteome, accelerating biological discovery through multi-omics analysis"); also, the final sentence of that section is rather confusing.

- comparisons with known regulators and regulated genes are only done via statistical tests/ enrichment and it woud be interesting to compare possible composition of extracted modules as sets of genes. It would thus be interesting to see if this method could capture regulatory modules that could be used for predictive tasks.

- the code in github is only partial; the full reproducibility of the results, although possible in theory is made harder since the scripts run to generate the results are not made available.

Also, some minor aspects:

- in the Introduction (3rd paragraph; sentence starting with "Previously, ..." it is not clear that when describing previous work that it comes from ref 36, not cited there; given the importance of this previous work, it should be more clear.

Finally, I note that since the study works over the same organism, it is "known" that the regulatory modules should be shared (while we may not know all of them) and are in fact underlying the generation of all these datasets. So, it would be surprising to find that the modules retrieved from these datasets would not contain some similarities. So, in a way, what would be surprising would be the lack of this consistency.

**********

Have all data underlying the figures and results presented in the manuscript been provided?

Large-scale datasets should be made available via a public repository as described in the PLOS Computational Biology data availability policy, and numerical data that underlies graphs or summary statistics should be provided in spreadsheet form as supporting information.

Reviewer #1: Yes

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer#2:  Summary of the paper:

The authors have used five independently produced datasets of RNAseq and microarray, and performed an ICA analysis on them to find underlying consistent patterns in latent space. Each independent component consists of weights of genes, and the few genes with a significant weight are grouped as iModulons. The iModulons were then characterized as regulatory, functional, genomic, or uncharacterized. Some of these iModulons were found in all datasets, and many were found in at least two, indicating that this technique can find relevant features of genetic data.

Comments:

Major

My main comment is novelty, the authors use same technique/same philosophy published recently (ref 26) and similar research design, i.e. multiple cancer datasets comparison, ref 35, so message Matrix factorization recovers consistent regulatory signals from disparate datasets is not new. I would recommend to authors focusing on biological story, otherwise from low-rank matrices is pretty clear that they are the similar between biological experiments, especially when dealing with linear techniques such as ICA. Also although authors did analysis on log-transformed data, I would like to see it this analysis holds true on standardized data (mean-cantered and scaled to unit variance), otherwise ICA is sensitive to scale of data and genes that are highly expressed in one dataset are also same genes that are highly expressed in another. Otherwise, technically paper sounds well and nicely written with nice graphics.

Minor

I recommend to change short tittle “iModulons are conserved across disparate datasets” - iModulons is not something established, neither informative.

Figure Files:

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org.

Data Requirements:

Please note that, as a condition of publication, PLOS' data policy requires that you make available all data used to draw the conclusions outlined in your manuscript. Data must be deposited in an appropriate repository, included within the body of the manuscript, or uploaded as supporting information. This includes all numerical values that were used to generate graphs, histograms etc.. For an example in PLOS Biology see here: http://www.plosbiology.org/article/info%3Adoi%2F10.1371%2Fjournal.pbio.1001908#s5.

Reproducibility:

To enhance the reproducibility of your results, PLOS recommends that you deposit laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions, please see http://journals.plos.org/compbiol/s/submission-guidelines#loc-materials-and-methods

Dear Prof. Palsson,

We are pleased to inform you that your manuscript 'Independent component analysis recovers consistent regulatory signals from disparate datasets' has been provisionally accepted for publication in PLOS Computational Biology.

Before your manuscript can be formally accepted you will need to complete some formatting changes, which you will receive in a follow up email. A member of our team will be in touch with a set of requests.

Please note that your manuscript will not be scheduled for publication until you have made the required changes, so a swift response is appreciated.

IMPORTANT: The editorial review process is now complete. PLOS will only permit corrections to spelling, formatting or significant scientific errors from this point onwards. Requests for major changes, or any which affect the scientific understanding of your work, will cause delays to the publication date of your manuscript.

Should you, your institution's press office or the journal office choose to press release your paper, you will automatically be opted out of early publication. We ask that you notify us now if you or your institution is planning to press release the article. All press must be co-ordinated with PLOS.

Thank you again for supporting Open Access publishing; we are looking forward to publishing your work in PLOS Computational Biology. 

Best regards,

Kiran Raosaheb Patil, Ph.D.

Associate Editor

PLOS Computational Biology

Weixiong Zhang

Deputy Editor

PLOS Computational Biology

***********************************************************

Reviewer's Responses to Questions

Comments to the Authors:

Please note here if the review is uploaded as an attachment.

Reviewer #2: The revised manuscript carefully takes into account the comments I raised previously. It appears to be now appropriate for publication in Plos Computational Biology

**********

Have all data underlying the figures and results presented in the manuscript been provided?

Large-scale datasets should be made available via a public repository as described in the PLOS Computational Biology data availability policy, and numerical data that underlies graphs or summary statistics should be provided in spreadsheet form as supporting information.

Reviewer #2: Yes

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No