Hello,
I have a VCF file on which I first use VariantFiltration
(GATK v3.7) with various filter expressions, such as "QD < 2.0"
and "MQ < 40.0"
, as well as various genotype filter expressions, such as "DP < 10"
and "GQ < 20"
.
I then use SelectVariants
with --excludeFiltered
and --excludeNonVariants
, but also with --setFilteredGtToNocall
, --maxFractionFilteredGenotypes 0.300000
and --maxNOCALLfraction 0.300000
.
However, in the end, the output VCF file still contains a site with more than 30% missing genotypes.
To discard such a site, I have to run SelectVariants
twice, the second time only with --maxNOCALLfraction 0.300000
.
It hence looks like --setFilteredGtToNocall
and --maxNOCALLfraction
have to be used sequentially, in two different runs of SelectVariants
. Am I right?
Best regards,
Timothée