All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
Thank you for taking all reviews seriously, especially for addressing the comparison with other databases, and for even adding more analysis.
Although the requested changes are minor, they are essential to improve the manuscript.
The major suggestion is to compare to similar databases or those with similar scope. The other important change suggested by the reviewers is to improve the database website, especially by revising the disease list + to add an exclusion filter.
Other useful suggestions, made by Reviewer 2, should also improve the methodology.
Finally, while the manuscript is readable and clear, some polishing and final language editing is needed— (I'm sure the authors can do that after implementing the suggested changes).
[]
No Comment
No comment
Regarding the use of databases for miRNA targets, the authors relied on the HMMD database that provides data about experimentally validated miRNA targets. The web-based tool described by the authors can include a filter that allow users to exclude miRNA targets based on the experimental method used to validate those miRNA targets. (For example excluding targets validated by methods that used mRNA expression analysis by QPCR and include only targets validated on the protein level, or based on the knockout/overexpression methods for functional analysis of those miRNAs…etc). This is due to the fact that validating miRNA targets biologically can be challenging.
The manuscript by Liu et al. describes a database, miRDRN, that compiles associations between miRNA, target genes/proteins, and diseases. The database considers multiple levels of data and computes correlation between miRNA, genes and diseases by using principles of the set theory. The data used to build miRDRN was obtained from other databases that mostly rely on experimentally-validated sources, however, the authors do not provide any comparison to previously published alike databases or even to the ones from which they obtained the data for building miRDRN (e.g. HMDD, TarBase).
-- line 111: how did the authors integrate the different data types ? or did they mean 'compiled' ?
-- the definition of Jaccard score for detecting the overlap between two sets of genes as explained by the authors is very informative and very well-suited to GO. However, for pathways and since the authors account for primary and secondary interactions, it will be worth weighing the genes/proteins in a pathway depending on how near/far they are from the target genes (i.e. order of steps from/to target in a pathway)
-- line 161: step 3 mentions that targets of a specific miRNA are collected from two sources (HMDD and TarBase). Does the authors take the overlapping targets from the two sources (intersection) or union ?
-- It is not clear to me how the authors integrated the miRNA-disease association data. They have elaborated on how miRNA-target associations were computed but not for the disease association. Also, it is not clear how they define co-morbidity from a computational point of view.
-- lines 292-293: Are the 26 targets that have literature support well-validated targets ? if yes, I would suggest that the authors curate a list of positive and negative association 'gold standard' targets and then use this set to quantitatively evaluate the performance of their tool.
-- The authors should provide a comparison to existing similar tools and highlight the advantages of using their tool in particular. I am not up-to-date on microRNA databases but a quick search shows there are several against which the authors should mention (e.g. HMDD, miRwayDB). Also several of these databases have experimentally validated targets, so I strongly suggest that the authors provide quantitative performance evaluation measures (AUROC, F1 score, FPR ... etc)
-- the authors should clearly hint that the findings form their tool are exploratory in nature with the goal of prioritizing miRNA-target-disease associations.
-- the discussion on potential failure of the anti-AD drug (case #3) is very interesting and demonstrates the utility of the database in full. I would strongly suggest that the authors follow the same for the other two case studies (cases #1, #2). The authors could also hint to potential applications of building similar databases in the introduction and highlight that their databases is very comprehensive in guiding associations across several types of data that could potentially lead to identifying new drug targets, understanding drugs mode of action ...etc.
minor comments:
-- the names of the diseases in the 'disease' drop box need to be revised, for instance 'acute', 'chronic' and 'hepatocellular' are not proper diseases names.
-- what is the default p value cutoff ?
-- line 132: the intersection and union symbols are reversed.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.