Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors of this outstanding manuscript on fundamental aspects of cancer biology are computational cancer biologists. They focus of cancer cell dormancy reflecting Drug-tolerant persisters as well as polyploid/multinucleated giant cancer cells (PGCCs). Why is this outstanding? Because these important therapy responses, that underlie resistance and tumor repopulation, are totally ignored in numerous (thousands) of papers by pioneering oncologists who propose cell "death" (e.g.,  treacherous apoptosis) stimulating anticancer strageties.  

I have no concerns if the manuscript is published as written. My suggestions below are totally optional, and if the authors decide to disregard them its totally fine with me.  

I don’t follow the sentence on lines 301 and 302. Specifically, the phrase “graded progression along a quiescence-senescence continuum.” Senescence is a form of reversible dormancy with overlapping features with PGCCs. For example, please see PMID: 34695579, an Editorial by three pioneering PGCC biologists: Jinsong Liu, Jekaterina Erenpreisa, and Ewa Sikora. (Perhaps the authors want to cite this Editorial to give credit to the remarkable contribution of these three individuals to the field of dormancy.)

I did not see the terms “amitosis” and “neosis” in this manuscript. This is an important mean by which PGCCs generate tumor-repopulating progeny, as demonstrated by Kirsten Walen and Rengaswami Rajaraman (and many others) since the early 2000s. (For details, please PubMed search for Walen, Rajaramana and amitosis).

How is therapy-induced dormancy impact the interpretation of radiosensitivity and chemosensitivity data obtained by 96-well plate cell proliferation/viability assays (e.g., the “MTT test”), the “gold standard” radio/chemosensitivity (colony formation) assay, and  tumor growth studies in live animals? Do all theses assays measure proliferation arrest (dormancy) and not death? If so, does cytotoxic anticancer treatment  convert proliferating cancer cells (deadly) to dormant cancer cells (even more deadly) and not dead cancer cells? Unfortunately the answer to all of these questions is yes!

The authors highlight heterogeneity. I would suggest being more specific. Does this refer to intratumor heterogeneity, with different subpopulations of cancer cells within the same tumor exhibiting resistance by different mutational and non-mutational (apoptosis-anastasis) mechanisms?

What is the role of apoptotic cancer cells in reawakening dormancy? Please PubMed search for “treacherous apoptosis” for the current knowledge on this.  

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

I appreciate the opportunity to review this manuscript. While the topic is highly relevant, the current composition functions more as a collection of isolated study summaries rather than a cohesive structural analysis, which limits its utility as a comprehensive review. There is a distinct lack of critical discussion regarding the correlations and contradictions between the cited studies. To improve the manuscript, the authors should move beyond mere description and instead synthesize the literature to highlight shared principles, conflicting findings, and future research directions, perhaps by elaborating on the conceptual framework presented in Figure 2. Currently, the depth of discussion is insufficient for experts and lacks the clarity required for newcomers. Furthermore, key molecular regulators like NRF2 are only briefly mentioned with minimal citations, which should be expanded to better inform the reader. Although the work appears scientifically sound, the structural limitations and the lack of analytical synthesis must be addressed to meet the expected standards for publication in this format.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This review focuses on computational biology research into cancer cell dormancy, systematically organizing the full spectrum of computational methods from dormant cell identification, microenvironmental regulation analysis, and state transition modelling to clinical prognosis association. It also accurately identifies core challenges in the field and proposes forward-looking solutions. The topic holds significant clinical and disciplinary value, with a clear logical framework and comprehensive, up-to-date literature coverage, making it a high-quality review manuscript. However, the manuscript still has areas requiring refinement, including insufficient depth in some content discussions, limited graphical support, and minor technical inconsistencies. It is suitable for publication after appropriate revisions.

 

(1) Insufficient discussion on the standardization of concepts and terminology

1. The manuscript points out the problem of confused term definitions and blurred subtype boundaries in cancer cell dormancy (e.g., overlaps between DTPs, DTCs, PGCCs, and SRCCs), but only mentions this in the introduction and challenges sections, without a dedicated subsection for systematic sorting and standardization recommendations of the existing terminology system. This may leave readers confused about the core characteristics and distinguishing criteria of different subtypes.
2. The conceptual analysis of the "dormancy continuum" vs. "discrete dormant states" is not in-depth enough. Briefly mentioned only in Section 3.3, it fails to compare the rationality of the two viewpoints with existing experimental and computational evidence, nor clarify the specific impact of this conceptual controversy on computational method design (e.g., applicable scenarios of clustering analysis vs. pseudotime analysis).

 

(2) Insufficient depth and detail in the discussion of some computational methods

1. Oversimplified methodological introduction: For key computational methods mentioned in the manuscript (e.g., NMF clustering, RNA velocity, pseudotime trajectory analysis, Latent ODE, MOFA+), only their applications in research are described, without a brief explanation of the core principles, applicable scenarios, and limitations of the methods. For readers without a computational biology background, it is difficult to understand the selection basis and result interpretation key points of different methods.
2. Lack of comparison and integration of different methods: There is no horizontal comparison of different computational methods for the same research question, such as the advantages and disadvantages of single-cell transcriptome analysis, lineage tracing, and spatial transcriptome analysis in "dormant cell identification"; the comparison of applicable data types and integration effects of methods such as MOFA+ and graph-linked unified embedding in "multi-omics integration". No reference framework for method selection is provided for researchers.
3. Insufficient discussion of the original tool LitGraph: As an AI-assisted literature mining tool developed by the authors’team, only its application results are introduced in Section 3.1, without explaining the tool’s core design ideas, technical workflow, and differences from existing literature mining tools, nor providing practical application cases (e.g., how to solve the problem of fragmented dormancy biomarkers). This weakens the innovation and reference value of this work.

 

(3) Inadequate graphical information support and room for optimization in visualization design

1. Few figures with low information density: Only Figure 1 and Figure 2 are mentioned in the manuscript, and the description of the figure content is brief. For example, Figure 1 only illustrates the formation mechanisms of DTPs/DTCs and the dormancy continuum, but does not clarify the core information of each module in the figure; the challenges and solutions in Figure 2 are only summarized in words, without clearly presenting the association of challenge-solution-corresponding computational method through visualization.
2. Lack of schematic diagrams for methodological workflows and cases: No schematic diagrams are drawn to assist understanding for the application workflows of key computational methods (e.g., lineage tracing combined with single-cell multi-omics to analyze dormancy state transitions, spatial transcriptome analysis for microenvironmental regulation), increasing the reading difficulty for non-computational background readers.
3. No summary tables: Core tables can be added to integrate key content in the manuscript, such as Core characteristics, inducing factors and computational research methods of each subtype of cancer cell dormancy, Application scenarios, advantages and limitations of different computational methods in cancer dormancy research, and Association between dormancy-related molecular biomarkers and clinical prognosis, to improve the readability and practicality of the review.

 

(4) Omissions and imbalance in the discussion of some research directions

1. Lack of cross-cancer comparison of dormancy characteristics: Most studies mentioned in the manuscript focus on breast cancer, lung cancer, and myeloma, without analyzing the differences in molecular characteristics, microenvironmental regulation, and dormancy dynamics of dormant cells across different cancer types, nor explaining the adaptability of computational methods in cross-cancer research.
2. Insufficient discussion of computational drug discovery targeting dormancy: The manuscript only mentions that mathematical modelling provides ideas for dormancy-targeted therapy in Section 2.4, but does not systematically sort out the application of computational methods in the identification of dormancy-related drug targets, drug molecule screening, and the design of combination therapy regimens, with insufficient integration with the research orientation of clinical translation.
3. Weak discussion of the combination of experiments and computation: The review focuses on computational methods but does not clarify how computational results guide experimental validation, nor mention how experimental research provides data support and parameter calibration for computational models. The discussion of the computation-experiment closed loop is lacking, which may cause readers to ignore the synergy between the two.

 

(6) Minor technical issues

1. Inconsistent literature citation formats: Some literature citations in the manuscript use superscript numbers (e.g., [1,2]), and others use author + year (e.g., Blath et al [60]), which need to be unified.
2. Unmarked first occurrence of some abbreviations: Although the first occurrence of "scRNA-seq" is marked in Section 2.1, the first occurrence of some niche abbreviations (e.g., "super-SILAC") is not explained in full, increasing reading difficulty.
3. Some repetitive statements: Core biomarkers of dormancy such as the ERK/p38 signalling pathway are repeatedly mentioned in the introduction, molecular mechanisms, computational applications and other sections without integrated discussion, leading to redundant content.

 

Comments on the Quality of English Language

Reconsider after major revision

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

I have confirmed the revised version and the authors' reply. Thank you very much.

Reviewer 3 Report

Comments and Suggestions for Authors

The author answered my questions very seriously. I have no further objections. This article is recommended for direct acceptance.