Allele flip in GATK
0
0
Entering edit mode
9 weeks ago
MAPK2 ▴ 40

I have this variant (rs17863783) from Visscher et al. 2013 , where they reported alleles as A,C for this variant.

However, when I checked this in my VCF, it has G,T. I was wondering if this is due to strand bias or allele flip in GATK and should be considered the same as A,C. Can someone please help me understand this?

 chr2       233693631       chr2:233693631:G:T      G       T

GATK VCF • 472 views
1
Entering edit mode

Hi! Doing a quick search it looks like the genotype of that SNP is G,T. Have you checked the Visscher et al. 2013 publication to see how they carried out the genotyping? Perhaps they employed a methodology based on complementary sequence? I'd take that direction; I doubt strand bias has anything to do with this :).

0
Entering edit mode

iraun Thank you. That seems a valid explanation. This is how they genotyped "DNA samples were genotyped using a custom 96-plex Illumina Veracode GoldenGate SNP genotyping assay according to manufacturer’s instructions (Illumina, San Diego, CA). This assay included an additional 63 non-study SNPs used for quality control purposes only". My guess is that GATK uses StrandOddsRatio or Fisher's Exact Test to evaluate strand bias and therefore the genotype by GATK is more valid.

1
Entering edit mode

No problem MAPK2 , happy to help. I am not an expert in that genotyping technology, but I just found this GoldenGate Genotyping Assay presentation and particularly the slide 6 seems to maybe clarify the issue?

0
Entering edit mode

You would need to check the original BAM but it could be that they called the variant on the opposite strand.