I think there are a few things that have to be clarified. Somatic variants are variants that arise in a certain tissue at a certain time point, but are not inherited by or from the parents via the germline, so oocyte or sperm. As a consequence, if one wants to perform somatic variant calling in e.g. a tumor biopsy, a normal control is required. That is most commonly a sequencing run from peripheral blood. In any case, it must come from the same patient. I wanted to point that out because you asking for any hg19 normal BAM will not help. It must be a sequencing experiment from the same source as your somatic sample. What kind of data do you have?
The command I use with the latest VarScan2 version is:
samtools mpileup -q 20 -Q 25 -B -d 1000 -f genome.fasta normal.bam tumor.bam | $varscan2 somatic /dev/stdin $outputname -mpileup --strand-filter 1 --output-vcf
This will create a combined tumor/normal pileup, piped into varscan, because there is not really a reason to save these pileups to disk I think.
modified 6 months ago
6 months ago by
ATpoint ♦ 14k