I am looking at somatic mutations across multiple samples, but the samples are not covered equally in many regions. Since we care about comparing the mutation counts between these samples, I would need to consider the uneven coverage -- e.g. sample A has 5X coverage at position 1, while sample B has 20X coverage at position 1; if I set filtering criteria about coverage in my workflow, and get rid of <10X mutations, then even if sample A has mutation in position 1, I would miss it; thus the comparison would not be fair.
Now my questions is, is there an easy/fast way to extract the well-covered regions across multiple samples? These are all WGS data (bam files size ~50-60GB), so I guess I could run bedtools on all of them and then overlap? Any other suggestions please? Thank you.
If you're afraid of missing such mutation, why do you need to extract well-covered regions ?
Most likely, you don't need to worry about coverage beforehand. VCF files include depth information, and often times, information about per-sample depth and / or per allele. So you can filter by coverage after variant calling.