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.
Since the remaining reviewer considers that the revision has been done (to some extent), I would like to recommend its acceptance to the journal. Congratulations!
[# PeerJ Staff Note - this decision was reviewed and approved by Jun Chen, a PeerJ Section Editor covering this Section #]
NA
NA
I've checked the authors' github space
https://github.com/wonderful1/CNV-P
and confirmed that they successfully addressed the point of concern.
NA
The revised manuscript has been reviewed by the two original reviewers. As you can see from their comments below, one of them is satisfied with your revision while the other points out that a few concerns remain, Please read the comments carefully and re-revise the manuscript accordingly unless you think that the comments are inappropriate.
All my questions are addressed.
All my questions are addressed.
All my questions are addressed.
The authors appropriately addressed my concern
no comment
The authors mostly addressed my concern, this reviewer appreciated it. However, in my original comment, I also asked the authors to compare with Werling et al. method
(now merged to) https://github.com/broadinstitute/gatk-sv#gather-sample-evidence
Please mention it and compare, or if there is difficulty to compare with it, please explain.
Also this reviewer would request the authors to describe the comparison pipeline at the authors' github space or somewhere, for the readers to be able to replicate it.
Your manuscript has been reviewed by two experts in the field. As shown in their comments below, both of them basically admit the value of this work while raising several important points; particularly, both request more detailed description in the main text. Please read their comments carefully and revise the manuscript accordingly. Looking forward to your revised manuscript.
[# PeerJ Staff Note: Please ensure that all review and editorial comments are addressed in a response letter and any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate. It is a common mistake to address reviewer questions in the response letter but not in the revised manuscript. If a reviewer raised a question then your readers will probably have the same question so you should ensure that the manuscript can stand alone without the response letter. Directions on how to prepare a response letter can be found at: https://peerj.com/benefits/academic-rebuttal-letters/ #]
In this manuscript, Wang et al. proposed a machine learning framework, CNV-P, for predicting high confident copy number variations (CNV). The computational detection of CNVs using genome sequencing data is an important task and should be of general interest to the community. I think this work may be helpful to the CNV data end-users.
The current CNV calling methods are suffering from low accuracy with high false positive rates. Different from the conventional methods which use hard cutoff to select high confidence callings, the authors proposed a machine learning method based on random forest to predict high confidence CNVs using multiple features such as read depth, split reads and read pair around the putative CNV fragments.
The authors validated their method on synthesized and real data sets. The proposed method showed high precisions with good sensitivity.
Minor Comment:
It is interesting to know the contributions by each feature in the random forest prediction model, and compared performance with the hard cutoff method using the eminent feature.
The detailed method of training in real sample is understandable by Table S1 and S2 but it should be written in the method section in the main text with similar details.
no comment
Integrate SV calls from multiple algorithms and extract better calls by using supporting evidences based on random forest learning of PE, SR, RD and so on, is not a new idea (for example, Werling et al. Nat Genet 2018; 50: 727-36. or Zhuang et al. NAR genom bioinform 2020; 2: Iqaa071, possibly more?) While the details were somehow different each other.
This reviewer asks the authors to compare the performance. It would also be valuable to see the comparison of the performance with other integrative CV callers such as MetaSV.
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.