Triallelic variant - GWAS

Question

[khn]

Hello,

If there is a SNP that surpassed GWAS significance, but it turned out to be triallelic, should we need to exclude this variant? I merged the two datasets, cases and controls, but no issues in the process to analyze in plink. But dbSNP says it is triallelic. In the map files for the two datasets, the SNP's alleles are both A/C. But dbSNP says it is A/C/T (T is ancestral).

Thank you in advance!

Answer 1

[chrchang523]

Do you know how the genotypes were initially measured? If the third allele could be expected to be systematically mistaken for one of the first two, you should adjust your interpretation of the result accordingly. But if the third allele legitimately doesn’t seem to be in your dataset and is generally rare in the population you’re studying, I wouldn’t worry about it.