inconsistency between somatic SNV from Varscan and SNV from Samtools

Question

[dingxm]

HI all

I have two paired samples. One is Normal and the other is Tumor. I called SNP of both samples using Samtools/bcftools and then call somatic mutation using Varscan2. I get some somatic mutation from Varscan. But these mutation are not among the mutation list of tumor sample I got from samtools. Can someone help to explain.

Best

Answer 1

[Sean Davis]

The two variant callers use very different theory for calling variants, so it is not surprising that the two give different results. It turns out that VarScan is very sensitive (or noisy, depending on your point-of-view) while samtools is less noisy. In practice, VarScan often needs to be further filtered to get reliable variants.

[dingxm]

Sean Thanks for your reply. These somatic mutations were got after strict filter. Is that possible these somatic mutations have low allele frequency and was not called as germline SNP by Samtools?

[Sean Davis]

Yes.