Read more about it here!
BQSR only uses position info from true SNPs vcf?
I'm currently using BQSR with a published dataset of SNPs to recalibrate my base quality scores. However some of the known loci are multiallelic and the vcf can't pass validation because the validation expects at most biallelic loci. I transformed these loci into missing data (./.) to keep the variant position in the vcf file, so that these positions would still be considered when performing the statistics in BQSR.
However, can you confirm if the base quality recalibrator is only using the positions (Chr#:position) from the dabase SNP vcf (and not using the genotype information)?