Hi @bhandsaker
The results of CNVdiscovery followed by compute allele frequencies, are a bit weird for me. I've attached two IGV images, showing both the genomestrip vcf file, and bam files from two of the samples.
CNVs, coloured by genotype, (red homozygous mutation, CN=4, green=heterozygous, CN=3).
In the first image, genomestrip identifies five events in the region show. Judging by the insert sizes, (and a bit of eyeballing the read depth) it's not five events but one tandem duplication, which is homozygous in the parent (lower) and heterozygous in the offspring.
The second image shows another duplication, which genomestrip has basically split into two.
I guess my problem is that because the read depth varies a lot across short genomic regions, genomestrip thinks these are CNVs. My solution would be to disable the read depth-based detection of CNVs, leaving just insert size and split reads to discovery CNVs.
Thus my question is: How do I disable the read-depth based detection?
Sincerely,
William Gilks