Open danielben-isvy opened 6 months ago
All INFO
fields in the final case VCFs are actually added by the GLIMPSE imputation step and I would hope the descriptions of the RAF
/AF
fields given are accurate.
Do we need to recalculate AF
for single-sample VCFs or can we take care of this after merging to a cohort VCF?
It's probably easiest to just recalculate after merging to a cohort VCF.
In the single-sample VCFs produced by the KAGE pipeline, the AF info field is calculated based on genotype dosage (DS) rather than the called genotype (GT), resulting in single-sample AFs that do not match the expected 0, 0.5, or 1.