We think that these modifications are necessary to comply with the VCF format of the GTEX data (with GT in the 2nd field, the first field is the GL). GL/GT/DS. However, maybe we should check if these will work with all VCF files?
"GT" isn't always at the 2nd position (see official spec). But the current code still has to be fixed. Can you try to add code that (1) find the location at which "GT" is in the column "FORMAT", and then (2) split the column and retrieve the genotype?
We think that these modifications are necessary to comply with the VCF format of the GTEX data (with GT in the 2nd field, the first field is the GL). GL/GT/DS. However, maybe we should check if these will work with all VCF files?