Merge vcf files from the four variant callers from TCGA

Question

[j.lunger18]

I'm interested in taking .vcf files obtained from TCGA and merging them into a single file for a group of the same samples. TCGA used 4 variant callers (Somatic Sniper, MuSE, MuTect2, and VarScan2). Is there a script that can do this?

Answer 1

[Kevin Blighe]

If you have access to the VCFs, then I assume that you have access to the BAMs? My preference, time-permitting, would be to re-call somatic variants from the BAMs using the same variant caller, but I'm aware that this may pose a logistical challenge (in addition to taking more time).

In any case, you can do this yourself by setting rules about which variants to keep (or not). Questions you need to ask (and answer):

1. are all samples profiled by each method? If not, which samples are profiled across more than one method?
2. if a sample has results across just 2 methods, should I only retain variants identified by both methods, or just take a Union of the variants identified by both?

BCFtools will be your friend for getting through this, and awk.

Kevin