Understanding the SARS-CoV-2–Human Liver Interactome Using a Comprehensive Analysis of the Individual Virus–Host Interactions

Round 1

Reviewer 1 Report

This paper by Giovanni Colonna is a very interesting, extremely rich in information brought by the analyzes which were carried out in an orderly and methodical way. It is not clear whether any mutations (at least known ones) of the virus are taken into consideration in the various analyzes with the various software/databases. In this case, is it possible to evaluate the pathological effect of a certain mutation of the virus? Is it possible in these analyzes to take into consideration the possible effect of possible mutations of the target proteins?

Author Response

REPORT 1

Dear Reviewer no. 1

Thank you for your comments

Below are my answers to the questions you asked me.

It is not clear whether any mutations (at least known ones) of the virus are taken into consideration in the various analyzes with the various software/databases. In this case, is it possible to evaluate the pathological effect of a certain mutation of the virus? Is it possible in these analyzes to take into consideration the possible effect of possible mutations of the target proteins?

The most studied mutations of the virus are those relating to SPIKE and the proteins involved in the virus-human cell interaction mechanism, because of their undisputed scientific interest. The effects of mutations on SARS NSPs or accessory proteins are much less studied. However, in order to highlight the relational effects of a protein that carries mutations in a network, researchers must experimentally validate its interactions with the human proteome, just like BioGRID does, by using the protein as physical bait. At the moment, this platform does not report interactions for mutated SARS-CoV-2 proteins. Added to this is another problem. STRING, which is the best platform for interatomic analysis, uses to collapse (it mentions it) onto the native protein any information on the functional characteristics of a mutated protein or its proteoforms (metadata). This peculiarity does not allow us to study a protein with different functions because its effect is to expand the relational capabilities that are associated with the native protein, giving it greater connectivity. It is necessary to highlight and control this alteration. When calculating networks, STRING generates a context-specific knowledge base by thoroughly analyzing scientific literature (PubMed) for each network model. Instead, other databases curate proteins according to the specificity of the database itself, making it difficult to use proteins from different databases together. Also, in this case, it is necessary to check, but it is more difficult without an automatic analysis of the literature with the help of AI, as STRING does.

If we insert the relational characteristics (structural and functional) of the mutated proteins into the databases, treating them as proteins independent of the native one, then we can certainly get metabolic information related to their functioning. Especially the ones affecting the inter-individual drug response and efficacy by altering binding affinity. The scientific community has studied these effects using datasets that are very limited, but we need an experimental basis with a large dataset of binding affinity changes because of single-nucleotide polymorphisms (SNPs) at binding sites in order to properly evaluate them. However, to the best of my knowledge, such a dataset does not exist.

However, there are first attempts to solve the problem by a molecular docking approach and using a multithreaded (multi-channel) virtual screening workflow for generate protein–ligand complexes. However, methods of biophysics have not experimentally proven the interactions or biochemical functional tests have not tested the functions. Despite having collected about 78,000 annotations of experimental evidence from many organisms where mutations have been shown to affect a protein interaction, also the new IntAct platform remains very limited.

Reviewer 2 Report

The paper is presenting an analysis of hub-nodes specific for the liver during covid-19, based on the information reported in 18 articles (2021-2023) from PubMed to obtain deep functional information on viral strategies.

It shows that over one half of the liver proteins are directly involved in Covid, an impairment of the ribosome and spliceosome, an antiviral defense mechanism against cellular stress of the p53 system. The virus is capable of multiple protein attacks against single human proteins, so that the authors assume underlying evolutionary-topological molecular mechanisms. The virus seems to have different molecular strategies for different organs.   

It is in my opinion adequately covering the subject, being very detailed in the description of methods and results.

The structure of the manuscript could in my opinion be improved, as it provides very detailed information which might be difficult to understand for non-specialists. 

The cited references are relevant to the topic. The figures are clear and easy to understand. 

I find the conclusions are consistent with the argumentation; they are supported by the listed citations. 

The work is offering an insight on the pathological mechanisms of viral infection and could help understanding the complexity of viral SARS-CoV-2 infection and related diseases, contributing to developing treatment strategies.  

Author Response

REPORT N.2

Dear Reviewer N. 2

Thank you for your comments

Below is my response to your concern.

The structure of the manuscript could in my opinion be improved, as it provides very detailed information which might be difficult to understand for non-specialists. 

I acknowledge your concern, but I had to address requests from another referee to shorten the manuscript, which I have completed. I included a “Conclusion” paragraph that specifically addresses a major problem in the research I presented. Researchers frequently overlook experimental controls when using indirect methods and we cannot identify genes as HUBs in subsequent analysis. While writing this manuscript, I observed how prevalent this distortion is, tainting our understanding of the field. While career pressure may drive us to publish quickly, it is crucial to honor Gallileo's teachings, as our scientific knowledge relies on experimentation. I'm hoping you share the same viewpoint as me.

Reviewer 3 Report

The article by Giovanni Colonna entitled “Understanding the SARS-CoV-2-human liver interactome using a comprehensive analysis of the individual virus-host interactions” identified patterns of non-random connectivity or a viral organizational strategy that often remains hidden when analyzing the many molecular details of biological graphs in case of SARS-CoV-2. It is an interesting article highlighting the important aspect of host virus interaction.

I have following concerns about the study.

1)      Article is scientifically sound; however, it is written very poorly. The article needs extensive English language editing.

2)      Some portion in the introduction section and material and methods section are unnecessarily lengthy, it needs to be more focused and precise. For example, subsection Biogrid, STRING can be more precise and shortened.

3)      Figure-1 is not visible, author need to provide enhanced quality of figure.

4)      Author needs to include conclusion in the discussion section.

Extensive editing of English language is required.

Author Response

REPORT 3

Dear Reviewer n.3

Thank you for your useful comments.

Below are responses to your concerns.

The structure of the manuscript could in my opinion be improved, as it provides very detailed information which might be difficult to understand for non-specialists. Article is scientifically sound; however, it is written very poorly. The article needs extensive English language editing.

Some portion in the introduction section and material and methods section are unnecessarily lengthy, it needs to be more focused and precise. For example, subsection Biogrid, STRING can be more precise and shortened.

I revised and condensed the introduction and subsections of STRING and BioGRID. I revised English throughout the manuscript.

Figure-1 is not visible, author needs to provide enhanced quality of figure.

 I got the image from STRING, which allows for its download. I downloaded a new, more detailed image (from 1.9 Kb to 3.9 Kb) but the poor definition is because of the resizing of the image on the page. Using the enlargements of the Word or Adobe pages, the names of the nodes can be clearly read.

Author needs to include conclusion in the discussion section.

The Discussion now includes a Conclusion paragraph. The second part of the Conclusion emphasizes the primary challenge I encountered in this study, which I believe the scientific community should consider. Experimental control is lacking in using indirect methods to study interactions. Closer analysis revealed that researchers did not verify many functional hypotheses with physical or functional methods. The literature incorporates these hypotheses and they can sometimes lead to distorted conclusions if we’re not careful. I hope we agree regarding my standpoint.