too many indels output by varscan2

[mrfox]

Hi all,

I used varscan2.3.2 to study groups of normal/tumor matched data sets. I used java -jar Varscan2.3.2 somatic [normal pileup] [tumor pileup]. It outputs about 9,000 indels, of which most of them are reported as Germline and about one hundred of them were LOH or Somatic.

This is my first exome-sequencing project. Does any one has an idea what happened to predict so many indels? BTW, I used bwa to map the reads and duplicates were removed before SNV call.

Thank you.

[dkoboldt]

Hello,

I'm not sure what you expected to get, but those numbers (~9,000 called, of which 99% were germline/inherited, and 1% were somatic or LOH) are fairly typical for tumor-normal comparisons. Note that both "Germline" and "LOH" variants are present in the normal, but the latter became homozygous in the tumor due to loss of one allele.

Note that you can use the processSomatic command of VarScan to parse your output and break it down according to somatic status (germline, LOH, somatic) and confidence (high or low).

[dkoboldt]

SeqAnswers is OK because we do monitor it, however, you might hear back faster from us if you post to the VarScan Help Forum.