Dear colleagues,

If you wanted to be sure that VCF from exome tumour-normal sequencing, focused on somatic variants you got from someone was done properly and you would believe BAM preparation and data quality, what would you check?

I would go for:

• in baits filter
• QUAL filter
• indels in homopolymer
• amount of reads supporting alt variant
• % of reads supporting alt variant
• 0 alt reads in normal tissue control
• min coverage
• max coverage (2x median)
• low complexity regions

Filtering by databases: - ExAC - SIFT - PolyPhen

What would you add? I can't find a list of existing filters anywhere. Do you have some favourite paper where they have done really precise filtering?

Thank you very much, Martina