We are running Mutect2 on WGS prostate cancer samples. We analyzed our tumor-normal pairs previously with GATK 126.96.36.199 and validated the results against deep sequencing panel results (500-700x).
We are now running the samples again with GATK 188.8.131.52 and are getting significantly more false positives. Same exact command line, Gnomad resource, PoN, filtering, ... everything. Just the docker container is different: broadinstitute/gatk:184.108.40.206 instead of broadinstitute/gatk:220.127.116.11
Are we missing some new parameters that we should launch Mutect2 with, should we change something in the filtering of the results or why has the performance gone down dramatically between the versions?
Please sign in to leave a comment.