I am trying to select for the best bin size for calling CNVs. In one post, creator of CNVnator said "You can use option ‘-eval’ which will print average RD and standard deviation. We recommend to chose bin size such that the ratio is around 4." I am a little confused about the -eval option because I have to state a bin size in order to run it, which means I would have had to select for a bin size already.
Generally though, it looks like bin size is inversely proportional to read depth (30 bp bins for 100X cvg and 500 bp pins for 4-6X cnvg)
However, in the CNVnator paper, they also said the bin size should not be comparable to read length, so what should I do if I have whole genome sequencing data with read length around 150 bp and read depth around 80 (this is an estimation)? I am working with non-human data, and using CNVnator 0.3.2. Thanks!
