Hi, Dear all! I am new to GWAS, there exists one question made me sad. I wonder why do we need to remove multi-allele when we conduct a GWAS or NGS-sequencing analysis like WES. I have tried to find the answer in the Google. But I cannot get any useful information. I sincerely wish that you could give me the right reason. Thank you so so so much for that!
It seems to me that when dealing with low-coverage data like genotyping arrays, multi-allelic variants are more likely to be false signals. They might represent errors in variant calling procedures instead of real mutations. That, however, is not true with WES or WGS sequencing data when the whole genome is well sequenced, and rare variants (especially in large cohorts like biobank-level data) are much more likely to be detected. I'm currently working on WES association analysis, and I don't simply remove multi-allelic variants. I split them for further analyses instead.
I hope you find my experience helpful.
Best regards.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
Thank you sooooo much!! I already know that we will consider the multi-allele as false signals when we conduct GWAS by genotyping arrays. I am wondering how to treat the multi-allele when we use WGS and WES data. And if we decide to remove them, the reason is what I am puzzled with. After all, thank you soooo much!
A very useful tool to split multi-allelic variants is bcftools. If you simply want to remove them, PLINK will do fine.