I am using varscan2 (for matched N/T pairs) to call somatic SNVs. For filtering, i used script from varscan to filter high confidence somatic call. I used two scripts from varscan, and i get completely different results with very little overlap.
java -jar VarScan.v2.3.7.jar processSomatic $name
This will be two output files, namely $name.Somatic and $name.Somatic.hc (high confidence output )
java -jar VarScan.jar somaticFilter $name.Somatic --min-coverage 15 --min-reads2 5 --min-strands2 2 --p-value 0.001 --output-file $name.Somatic.hc"Criteria2"
Overlap between two high-confidence somatic calls,
overlap is about 10%. Why two filtering criteria are so discordant.
Any insight will be helpful.