Missing gentoypes despite high number of reads
GenotypeGVCFs, sets all genotypes as missing for a large part of a gene I am interested in, even though there seem to be lots of reads in all individuals for those sites. I have the same problem if I am using versions 3.5 or 3.7. It also makes no difference if HaplotypeCaller was run with -ERC BP_Resolution or the standard GVCF, or if I run GenotypeGVCFs on just one individual or many at once. Could it be related to the fact that this is a mitochondrial region and thus has higher sequencing depth and theoretically no heterozygote positions?
Here is an example of the vcf file produced by GenotypeGVCFs:
chrM 16758 . A . 67.16 . DP=14015 GT:AD:DP ./.:444,9:453 ./.:257,0:257 (...)
In the g.vcf file of the first individual this site looks like this (the 9 alternative allele reads are likely contamination):
chrM 16758 . A . . . GT:AD:DP:GQ:PL 0/0:444,9:453:0:0,0,0
And in the second individual g.vcf:
chrM 16758 . A . . . GT:AD:DP:GQ:PL 0/0:257,0:257:0:0,0,0
Any advice would be highly appreciated.