Hi, I'm just working on understanding VCFs and appropriate filters to use.
If the AD column is 10,0 but the genotype is called as 0/1, how much confidence can you have in this?, when no reads have been observed for the alternative allele. In my uninformed opinion this would be an easy filter step, remove all variants called as 0/1 but with no ALT reads observed but surfing around the GATK forum, seqanswers, etc, no one appears to be filtering by AD, why is this?
Many thanks.
UPDATE - this looks like it might be a bug in GATK, e.g. when a het is called with no reads for the alternative allele:
0/1;0,10
possible like the one recorded here - http://blog.goldenhelix.com/?p=1534
just doing some further checking.
UPDATE AGAIN - not a bug, just unexpected behaviour, see
http://gatkforums.broadinstitute.org...tions-of-reads
If the AD column is 10,0 but the genotype is called as 0/1, how much confidence can you have in this?, when no reads have been observed for the alternative allele. In my uninformed opinion this would be an easy filter step, remove all variants called as 0/1 but with no ALT reads observed but surfing around the GATK forum, seqanswers, etc, no one appears to be filtering by AD, why is this?
Many thanks.
UPDATE - this looks like it might be a bug in GATK, e.g. when a het is called with no reads for the alternative allele:
0/1;0,10
possible like the one recorded here - http://blog.goldenhelix.com/?p=1534
just doing some further checking.
UPDATE AGAIN - not a bug, just unexpected behaviour, see
http://gatkforums.broadinstitute.org...tions-of-reads