I have performed targeted Illumina sequencing of a particular gene of interest using UMIs, with processing according to GATK best practices. I have generated VCF files with Mutect2 and my goal is to look for somatic mutations.
When I look at processed BAM files in IGV, I see a high coverage in the 1000 - 20'000x range (depending on the batch; batches with more pooled samples had lower coverage). However, when I examine the VCF files, I see that the DP values are considerably lower than the read counts displayed in IGV. Several samples have DP < 100 (one has only got 7!), while others have DP in the 1000-5000 range. I'd like to have high DPs in order to maximise my sensitivity when looking for somatic mutations.
Is this a common issue? Can you suggest some ways to troubleshoot it, both sample prep-based (e.g. sequence more DNA, pool fewer samples) and analysis-based (e.g. more permissive quality filtering)?
show us the output of samtools flagstats in a region with such problem please