Completed on 13 Jun 2017 by Matthew Crook . Sourced from http://biorxiv.org/content/early/2017/02/11/107862.
Login to endorse this review.
A few things that I think would make the article stronger:
1. Compare your tool to PASC (https://www.ncbi.nlm.nih.go... and its accompanying paper (DOI 10.1007/s00705-014-2197-x). This would need to be addressed in the text and added to tab 8 of Supplemental File 1.
2. Address how VICTOR handles (or fails to adequately handle) known problems in the current ICTV classification (e.g. Enterobacteria virus lambda and Enterobacteria virus P22 are highly similar, but they are placed in Siphoviridae and Podoviridae, respectively—see DOI 10.1016/j.resmic.2008.04.005 and DOI 10.1046/j.1365-2958.2003.03580.x, for example).
3. VICTOR currently outputs .pdf files of the calculated trees, .phy files of the calculated trees, and .csv files of the recommendations for assigning the submitted genomes to the same or different species, genera, and families. This is rather opaque. It would be helpful if VICTOR also provided .csv files of all the pairwise distances and a .txt file of the thresholds chosen by OPTSIL. Even better would be if it also took the submitted genomes and assigned them to existing ICTV species, genera, and/or families (or to the new, unnamed clusters identified in this paper) when appropriate instead of merely assigning them a sequential number.
4. Provide all pairwise calculations for all genomes analyzed by the three formulas (D0, D4, and D6) as supplemental files and indicate the thresholds chosen by OPTSIL.
5. The first six pages of the .pdf for Supplemental Figure 5 are empty 25.85-inch × 0.04-inch pages. This needs to be fixed.