I have some bam files from cancer patients' tumor tissues and blood. The data were originally used for somatic variant calling. But I wonder if I can use the blood data to call germline variants using GATK HC tool.
If you are doing somatic variant calling, you are using a tumor and a normal sample. Normal sample is used for determining germline variants.
I don't see any reason why you can't.
Then if I want to get all SNPs for whole genome, can I do imputation after genotype detection?
I've never done anything like that, but I would expect that the SNP data you obtain from WES is insufficient for WGS imputation.
Technically you can. However there is a certain chance that blood cells are contaminated with circulating tumor cells. If you're confident about your normal purity, it shouldn't affect much.
Unless you are working with blood cancers, circulating tumor DNA fraction is extremely low.
Is there a way I can calculate normal purity from BAM files?
I am not sure of methods. But, see below comment by @igor circulating tumor cells are extremely low, so should be okay.
Login before adding your answer.
Use of this site constitutes acceptance of our User Agreement and Privacy