filtering SNV/INDEL in tumor-only WGS samples

Question

[Bogdan]

Dear all,

To the cancer genomics community on Biostars that worked with tumor-only samples, a question please : which criteria do you apply to identify the most reliable somatic mutations ? I would appreciate having your suggestions.

<> my colleagues performed somatic mutation calling with Strelka using as a reference genome the WGS of a cell line with normal phenotype NA12878

<> after excluding gnomAD and ExAC and dbSNP positions, we are left with ~ 100 000 mutations.

<> I am filtering these mutations by using AD (Allele Depth) or AF (Allele Fraction) criteria : for example, AD > 20 reads, and AF > 0.2

Beside these steps, are there any additional criteria that you would recommend ?

Thanks a lot.

Bogdan