stephenslab / susieR

R package for "sum of single effects" regression.
https://stephenslab.github.io/susieR
Other
182 stars 47 forks source link

Multi-allelic SNVs and CNVs #237

Open katherinef opened 2 months ago

katherinef commented 2 months ago

Hello,

I think Susie can only handle multi-allelic variants when these have been split into biallelic variants - is this correct? Unfortunately this isn't always easy to do (especially for CNV where you may not know what the underlying alleles are). Is there any way to use Susie with multiallelic SNVs and CNVs (ie, where the genotypes are coded 0,1,2,3,4 and so on instead of the usual 0,1,2)?

Thanks!

pcarbo commented 2 months ago

@katherinef Thanks for your interest in susieR. The short answer is that susie should be able to handle multi-allelic variants because it makes no specific assumptions about the input matrix, X. So it would depend entirely on how you encode the multiallelic SNV in the matrix X. I suspect that there are standard ways to do this; to start, maybe see how PLINK handles multiallelic SNVs.