Our vcf is created with samtools/bcftools mpileup. It has a parameter -L that sets a depth cutoff to disable INDEL calling (doc here but sadly was not included in doc for the version we use, prob got missed when mpileup moved from samtools to bcftools). Our data usually has much higher depth than the default cutoff, which means in most of our samples, INDEL calling is turned off and therefore don't show up in the vcf.
Consensus genome is unaffected, but its generation also uses samtools mpileup. So if to update samtools/bcftools version, testing the fix would need to cover both places.
Another outstanding issue in our vcf is here, don't see it very often but very curious.
Our vcf is created with
samtools/bcftools mpileup
. It has a parameter-L
that sets a depth cutoff to disable INDEL calling (doc here but sadly was not included in doc for the version we use, prob got missed whenmpileup
moved from samtools to bcftools). Our data usually has much higher depth than the default cutoff, which means in most of our samples, INDEL calling is turned off and therefore don't show up in the vcf.Consensus genome is unaffected, but its generation also uses
samtools mpileup
. So if to update samtools/bcftools version, testing the fix would need to cover both places.Another outstanding issue in our vcf is here, don't see it very often but very curious.
See also #56