DP and AD tags in VCF

[thedam]

Hey,
in .vcf there are many tags like this:

chr1 924310 . C G 49 PASS DP=5;ADF=0,5;ADR=0,12;AD=0,17; GT:DP:ADF:ADR:AD:BSD:BSQ:ALFR 1/1:5:0,5:0,12:0,17:0,0,0,5,12,0,0,0:0,0,0,39,39,0,0,0:0.000,3.400
chr1 924321 . C G 58 PASS DP=8;ADF=0,8;ADR=0,13;AD=0,21; GT:DP:ADF:ADR:AD:BSD:BSQ:ALFR 1/1:8:0,8:0,13:0,21:0,0,0,8,13,0,0,0:0,0,0,38,39,0,0,0:0.000,2.625
chr10 71390271 . C T 1000 PASS DP=33;ADF=0,0;ADR=6,27;AD=6,27; GT:DP:ADF:ADR:AD:BSD:BSQ:ALFR 0/1:33:0,0:6,27:6,27:0,0,32,0,0,27,6,0:0,0,38,0,0,39,39,0:0.182,0.818

so in short:

DP=5, AD=0,17   - incorrect
DP=8, AD=0,21   - incorrect
DP=33, AD=6,27 - correct

DP should be sum of ADs, yes? Why in some cases the values are wrong?

[thedam]

also such example:
chr1 779047 . G A 1000 PASS DP=15;ADF=0,15;ADR=0,0;AD=0,15 GT:DP:ADF:ADR:AD:BSD:BSQ:ALFR 0/1:15:0,15:0,0:0,15:15,0,0,0,0,0,0,42:39,0,0,0,0,0,0,39:0,1
Why GT=0/1, and AD=0,15? Shouldn't it be 1/1?

[bsiranosian]

Can confirm I'm seeing a lot of heterozygous calls that should be homozygous variant. If all the information is coming from the reverse strand, BS-Snper only calls homozygous variant if there are few to none reads coming from the forward strand, just by chance. I haven't looked at the source yet but this is probably a fixable bug.