search

Illumina / manta

Structural variant and indel caller for mapped sequencing data
GNU General Public License v3.0
398 stars 151 forks source link

Miscalling of HomRef? #189

Open weishwu opened 5 years ago

weishwu commented 5 years ago

I've run Manta on a WGS sample and one of the called variants was labelled as "HomRef" in the FT field in one sample:

chr1 2789490 MantaBND:0:14411:14412:0:7:0:0 T ]chr11:44096312]T 300 PASS SVTYPE=BND;MATEID=MantaBND:0:14411:14412:0:7:0:1;CIPOS=0,13;HOMLEN=13;HOMSEQ=CTTTCTTTCTTTC;BND_DEPTH=103;MATE_BND_DEPTH=63 GT:FT:GQ:PL:PR:SR 0/1:PASS:100:150,0,229:15,5:7,9 0/0:HomRef:26:24,0,112:6,0:6,3 0/1:PASS:42:250,0,39:4,4:0,12

In the second sample, GT is 0/0 and FT is HomRef, however PL is 24,0,112, and SR is 3 for the variant allele. There seems to be conflict between GT/FT and PL/SR. Which should I believe?

Thanks,

ctsa commented 5 years ago

The values are not in conflict -- the genotype assignments are not taken directly from the likelihoods (PLs)-- the GT and FT values here reflect the final genotype posterior, including the prior.