Hi,
I am calling variants in a family and I have observed that some variants are being called homozygous although they are heterozygous because the second allele is present only in a very small amount of reads (like 3%). I still know it should be heterozygous when for instance one parent is 0/0, the other is 0/1 and the child is 1/1. When I look into the mappings I can see that the ref allele of the child is only present in a very small amount of reads. How can I make it get called?
I can't set -stand_call_conf and -stand_emit_conf to lower values as they don't work in GVCF mode as far as I know. So far I have tried to use CalculateGenotypePosteriors and PhaseByTransmission but the wrong calls aren't being corrected. The mapping quality of the reads supporting the allele not being called is not to low. My ped files contain trios and should be fine.
Here's an example: (The first individual is the child which should be 0/1)
"gene 443 . A T 16238.90 . AC=3;AF=0.500;AN=6;BaseQRankSum=4.55;ClippingRankSum=-4.270e-01;DP=1116;ExcessHet=3.0103;FS=1.264;MLEAC=3;MLEAF=0.500;MQ=60.00;MQRankSum=1.83;QD=18.47;ReadPosRankSum=4.38;SOR=0.777 GT:AD:DP:GQ:JL:JP:PL:PP:TP 1/1:28,462:490:99:119:60:12212,708,0:12152,648,0:40 0/1:201,188:389:99:119:60:4063,0,4319:4123,0,4319:40 0/0:237,0:237:60:119:60:0,120,1800:0,60,1740:40"
I can see that the likelyhood for 0/1 for the child is very low but it is still supposed to be heterozygous. The allele frequency is probably the way it is due to sequencing or amplification. I read that putting the ploidy value up is one thing that can be done. Is there anything else? Because my ploidy is actually 2, it doesn't seem quite right to me to put it up.
Many Thanks!
Nadia