Dear Dr. Chiu,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Oct 01 2025 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org . When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Li Yang, M.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please note that PLOS ONE has specific guidelines on code sharing for submissions in which author-generated code underpins the findings in the manuscript. In these cases, we expect all author-generated code to be made available without restrictions upon publication of the work. Please review our guidelines at https://journals.plos.org/plosone/s/materials-and-software-sharing#loc-sharing-code and ensure that your code is shared in a way that follows best practice and facilitates reproducibility and reuse.

3. Thank you for stating the following in the Acknowledgments Section of your manuscript:

This work was supported by the grants from the National Science and Technology Council (NSTC).

We note that you have provided funding information that is not currently declared in your Funding Statement. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form.

Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows:

This research was supported by grants from the National Science and Technology Council (NSTC), Taiwan (NSTC 112-2222-E-130-003- and NSTC 113-2221-E-130-005-MY3), which assisted in the acquisition of computational facilities essential for this study.

Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

4. Thank you for stating the following in your Competing Interests section:

No available

Please complete your Competing Interests on the online submission form to state any Competing Interests. If you have no competing interests, please state "The authors have declared that no competing interests exist.", as detailed online in our guide for authors at http://journals.plos.org/plosone/s/submit-now

This information should be included in your cover letter; we will change the online submission form on your behalf.

5. In the online submission form, you indicated that your data will be submitted to a repository upon acceptance. We strongly recommend all authors deposit their data before acceptance, as the process can be lengthy and hold up publication timelines. Please note that, though access restrictions are acceptable now, your entire minimal dataset will need to be made freely accessible if your manuscript is accepted for publication. This policy applies to all data except where public deposition would breach compliance with the protocol approved by your research ethics board. If you are unable to adhere to our open data policy, please kindly revise your statement to explain your reasoning and we will seek the editor's input on an exemption.

6. Please amend your list of authors on the manuscript to ensure that each author is linked to an affiliation. Authors’ affiliations should reflect the institution where the work was done (if authors moved subsequently, you can also list the new affiliation stating “current affiliation:….” as necessary).

7. Please upload a new copy of Figure 1 as the detail is not clear. Please follow the link for more information: https://blogs.plos.org/plos/2019/06/looking-good-tips-for-creating-your-plos-figures-graphics/"" https://blogs.plos.org/plos/2019/06/looking-good-tips-for-creating-your-plos-figures-graphics/

8. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

9. If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise. 

Additional Editor Comments:

Thanks for submitting your work to PLOS ONE. Your manuscript has now been assessed by our editorial team and external peer experts. While they found it interesting, you will see that they have raised many serious problems and are advising that you revise your manuscript thoroughly. At the same time, please submit the point-by-point responses to reviewers' comments. If you are prepared to undertake the work required, I would be pleased to reconsider my decision. Please note that this revision decision does not assure the acceptance of your work. Thanks for the opportunity to consider your work.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Yes

Reviewer #2: Partly

Reviewer #3: Partly

Reviewer #4: Yes

Reviewer #5: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously? -->?>

Reviewer #1: Yes

Reviewer #2: No

Reviewer #3: I Don't Know

Reviewer #4: No

Reviewer #5: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available??>

The PLOS Data policy

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

Reviewer #5: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English??>

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

Reviewer #5: Yes

**********

Reviewer #1: The manuscript presents a well-structured computational framework that integrates single-cell RNA-seq deconvolution with weakly-supervised, attention-based multiple-instance learning to infer immune-cell composition directly from H&E whole-slide images (WSIs) of hepatocellular carcinoma. The work is timely, the methodology is technically sound, and the results demonstrate a clear improvement over the CLAM baseline. Nevertheless, several substantive issues—ranging from terminology precision to external validation—must be addressed before the paper can be considered for publication in PLOS ONE. I recommend Major Revision.

Major Comments

1. Throughout the manuscript the authors refer to “liver cancer” or “LIHC” without specifying that the study is restricted to HCC. Because intrahepatic cholangiocarcinoma and combined HCC-ICC have distinct immune microenvironments.

2. The Abstract and Discussion repeatedly label the approach as “non-invasive.” Since WSIs are derived from surgically resected or biopsied specimens, the technique is technically post-operative/ex vivo, not non-invasive.

3. The study relies solely on publicly available TCGA-LIHC WSIs and GEO scRNA-seq data. The manuscript currently lacks independent clinical validation (e.g., a local HCC cohort with flow-cytometry or multiplex-IHC ground truth). Please add prospective validation with HCC specimens.

4. Figure 5A (UMAP after log-normalization) is supplied at insufficient resolution. Please provide a high-resolution vector graphic.

5. Report exact p-values (not “p < 0.05”) for all log-rank tests in Supplementary Figure 10E–H and include 95 % confidence intervals for the accuracy metrics in Table 1.

6. Provide a brief justification for selecting log-normalization over CCA or SCTransform, despite the latter showing tighter batch correction.

Reviewer #2: The manuscript presents a timely and technically innovative framework (LIMPACAT) that integrates single-cell RNA-seq data and whole-slide imaging (WSI) to predict immune cell compositions in liver cancer. The multi-omics attention transformer approach is novel and shows promising results, particularly in stratifying patients by immune cell infiltration with implications for prognosis.

However, several important concerns should be addressed before the manuscript can be considered for publication:

1- Technical validity and data support:

The manuscript is partially supported by the data. While the use of deconvolution from scRNA-seq to simulate bulk RNA-seq is methodologically reasonable, there is no direct experimental validation of the immune cell predictions (e.g., via flow cytometry, IHC, or spatial transcriptomics). Moreover, all model training and evaluation were conducted using a single dataset (TCGA-LIHC), limiting generalizability.

2 -Statistical analysis:

Statistical approaches are broadly appropriate but lack sufficient rigor and transparency. There is no report of confidence intervals or statistical significance for model accuracy, nor of hazard ratios in survival analyses. No corrections for multiple hypothesis testing are mentioned, and data splitting or cross-validation strategies for model training are not described, raising concerns about potential overfitting.

3- Language and presentation:

The manuscript is not yet written in fully standard English. While the technical content is generally understandable, there are multiple grammatical and syntactic issues that require revision. Additionally, some sections (especially Methods and Discussion) are repetitive or overly verbose. A professional language edit is strongly recommended to improve clarity and readability.

In summary, this manuscript presents a creative and impactful contribution to digital pathology and computational immuno-oncology. With improved statistical rigor, clearer methodological reporting, and stronger language polishing, it has the potential to make a meaningful contribution to the field.

Reviewer #3: The authors present LIMPACAT (Liver Immune Microenvironment Prediction and Classification Attention Transformer), a framework that utilizes whole-slide images to predict immune cell compositions relevant to liver cancer prognosis. The concept is innovative; however, the manuscript is very difficult to follow in many key areas and it requires significant revision before it can be adequately evaluated.

Major Comments:

1. Clarity of Model Architecture and Workflow:

The inputs and outputs of the MIL-ATTENTION model are unclear in Figure 1 and throughout the manuscript. What is the output of the MIL-ATTENTION model? Figure 1 should also clearly distinguish between the training pipeline and the inference pipeline for both the ensemble DL-CCD model and the MIL-ATTENTION model, including the type of data and processing applied in each.

2. Positioning Against Existing Methods:

Many deconvolution models exist for bulk RNA-seq. The authors need to clearly articulate the advantages of their DL-CCD model over established methods.

3. Reorganization of the Results Section:

The sections from “Gene Count and Quality Control Metrics” to “Single-Cell Type Annotation” are too detailed for the Results section and should be condensed into a brief summary, with the full methodological descriptions moved to the Methods section. The Results should highlight findings that are novel and relevant to the field.

4. Validation Using Real Bulk RNA-seq Data:

The evaluation of the cell deconvolution model appears to rely primarily on simulated bulk RNA-seq derived from scRNA-seq. However, scRNA-seq and actual bulk RNA-seq have distinct characteristics. The model should be validated on real-world bulk RNA-seq datasets with known or estimated cell compositions to establish robustness.

5. Clarity and Framework Usability:

The section titled “Cell Composition Deconvolution Model and Immune Cell Level Estimated from WSI” is currently difficult to follow and lacks detail. It is the core of the paper and should be rewritten to clearly describe how the model components interact. It should also contain information about how the overall framework can be used by others.

6. Overfitting and Generalization:

There appears to be overfitting across all cell types. The authors should address this issue and describe any strategies used to mitigate overfitting (e.g., regularization, validation, dropout, etc.).

Minor Comments:

7. Acronyms such as “WSIs” and “CCD” should be spelled out at first mention. This is especially important in the abstract and background sections.

8. Several sentences in the Background section (e.g. paragraphs 2 and 3) lack citations.

9. Why using CCA and SCT normalization on LIHC bulk RNA-seq samples?

Reviewer #4: I read the submission titled “LIMPACAT:Multi-Omics Attention Transformer for Immune Prediction in Liver Cancer Using Whole-Slide Imaging” and thank the author for this interesting piece of original research. This manuscript presents a novel and technically compelling framework that integrates single-cell RNA-seq, simulated bulk transcriptomics, and transformer-based MIL models to predict immune composition from WSIs in HCC. The methodology is creative, well-motivated, and appears technically sound. Several important checks, such as the benchmarking of normalization methods and MIL model comparisons are included.

However, I believe the structure of the paper can be improved and additional information is necessary to make results reproducible and increase accessibility to a wider audience:

1. The methodology section is short and several important methodological steps are only introduced in the results. The first paragraphs of the results section do not contain any results and instead focus on quality control steps, software libraries and clustering methods. A clearer separation of methods and results would improve readability.

2. The data sets used are insufficiently characterized. I could not find information on the survival analysis data set, including number of patients that had matched WSIs and survival data available, their demographic and disease characteristics, the duration of follow up and whether there was any clustering present that could influence the statistical analysis (is there one slide per patient or potentially multiple?).

3. It is not entirely clear why only 20 samples were chosen from the GSE189903 dataset. This is generally a very small sample. Were the other 14 from different types of liver cancer (that would be a very high proportion) or was this number based on statistical or computational considerations? Do the 20 samples correspond to 20 different patients?

4. The validation using survival analysis is insufficiently described and could be more thorough. It is not clear how low and high CCD prediction groups were determined, where the cutoff was set and why CCD was not validated as a continuous predictor. Likewise, the paper mentions good and poor survival groups, but does not describe how they were determined. Using the raw prediction scores and survival times directly for validation would have been better. Instead of accuracy the C-index could have been reported. The method for splitting the sample into train/test or cross validation folds is not described. Uncertainty should be quantified. Alternative explanations for the reported associations should at least be discussed. A multivariate model adjusted for (at least) tumor stage could improve interpretability.

Reviewer #5: This study presents LIMPACAT, a novel and timely deep learning framework for predicting clinically relevant immune cell populations in liver cancer directly from whole-slide images. The approach of integrating single-cell and bulk transcriptomics with digital pathology is at the forefront of computational oncology. The findings that higher B and NK cell levels correlate with better prognosis are consistent with existing literature and demonstrate the potential of this tool for non-invasive biomarker discovery. While the work is promising, the following points should be addressed to strengthen the manuscript.

Major Revisions

1.Validation of Deconvolution Predictions: The most significant limitation of this study is the reliance on computationally inferred immune cell compositions as ground truth labels without subsequent biological validation. The model's high accuracy indicates it can successfully predict these inferred labels, but whether these labels accurately reflect the true cellular makeup of the TME is unknown. The authors should:

① More explicitly and prominently state this limitation in the Discussion and Conclusion sections.

② Suggest future work involving validation against a ground truth, such as performing immunohistochemistry (IHC) or multiplex immunofluorescence (mIF) for the predicted cell types (B cells, NK cells) on a subset of the WSI slides to correlate cell counts. For an example of how such validation is performed, the authors could refer to recent studies that correlate computational predictions with IHC.

2. Comparison with State-of-the-Art: The discussion compares LIMPACAT to several other models, including CLAM. However, the field of transformers in computational pathology is advancing rapidly. The manuscript would be significantly strengthened by discussing and contextualizing LIMPACAT with other very recent (2023-2025) transformer-based architectures for WSI analysis and immune prediction. This would provide a clearer picture of where LIMPACAT stands in the current landscape.

Minor Revisions

1. Improvement of figure quality:

The images of Supplementary Figure 10 (A)-(E) were exported out rather low resolution. Please ensure image clarity.

2.Methodological Clarity:

①In the Methods section, please specify the exact threshold or method used to stratify patients into "high" and "low" immune cell groups for the survival analysis (e.g., median, quartiles).

②Provide a more detailed description of the custom CNN, ATT, and ATT_TRANS model architectures, either in the main text or as a supplementary methods section.

**********

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: Yes:  Yigang Zhang

Reviewer #2: Yes:  VALERIA DUARTE DE ALMEIDA

Reviewer #3: No

Reviewer #4: No

Reviewer #5: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment 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. Registration is free. 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 PLOS at figures@plos.org . Please note that Supporting Information files do not need this step.

Dear Dr. Chiu,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Jan 10 2026 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org . When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Li Yang, M.D.

Academic Editor

PLOS ONE

Journal Requirements:

If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise. 

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

Additional Editor Comments:

Please further address the reviewer's minor concerns.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

Reviewer #1: All comments have been addressed

Reviewer #3: (No Response)

Reviewer #4: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions??>

Reviewer #1: Yes

Reviewer #3: Partly

Reviewer #4: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously? -->?>

Reviewer #1: Yes

Reviewer #3: I Don't Know

Reviewer #4: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available??>

The PLOS Data policy

Reviewer #1: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English??>

Reviewer #1: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

Reviewer #1: The authors have addressed all of my concerns in a detailed and rigorous manner. They have made the necessary revisions to the manuscript, and I believe that the current version of the article is now acceptable for publication.

Reviewer #3: Many places in the manuscript contain "Error! Reference source not found" that should be fixed.

Re Reviewer #3Response2, please point out the paragraph that "explicitly compare DL-CCD with established bulk RNA-seq deconvolution methods, such as quanTIseq and CIBERSORTx".

Reviewer #4: My comments were adequately addressed. I support acceptance given the additional inclusion of a separate validation cohort.

**********

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: Yes:  Yigang Zhang

Reviewer #3: No

Reviewer #4: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

To ensure your figures meet our technical requirements, please review our figure guidelines: https://journals.plos.org/plosone/s/figures 

You may also use PLOS’s free figure tool, NAAS, to help you prepare publication quality figures: https://journals.plos.org/plosone/s/figures#loc-tools-for-figure-preparation. 

NAAS will assess whether your figures meet our technical requirements by comparing each figure against our figure specifications.

LIMPACAT:Multi-Omics Attention Transformer for Immune Prediction in Liver Cancer Using Whole-Slide Imaging

PONE-D-25-22455R2

Dear Dr. Chiu,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager®  and clicking the ‘Update My Information' link at the top of the page. For questions related to billing, please contact billing support .

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Li Yang, M.D.

Academic Editor

PLOS One

Additional Editor Comments (optional):

Thanks for the authors' efforts to comprehensively improve your manuscript according to editor's and reviewers' comments. I am pleased to inform you that your paper can be accepted for publication now.

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

Reviewer #3: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions??>

Reviewer #3: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously? -->?>

Reviewer #3: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available??>

The PLOS Data policy

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English??>

Reviewer #3: Yes

**********

Reviewer #3: My comments were adequately addressed. The current version of the article is acceptable for publication.

**********

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 #3: No

**********