GWAS on a twin sample

Question

[javadjamshidi]

Hi everyone,

I want to conduct a GWAS on a twin sample, I was wondering if there is a method to conduct the GWAS on the whole sample without the need to remove the related individuals (twins). I know software such as GCTA and GEMMA can generate a GRM and use it as a random effect to fit a linear mix model, but I am not sure if it totally accounts for the relatedness in the sample.

Many thanks Javad

Answer 1

[Kevin Blighe]

There have been studies published on this topic - have you checked those? How many twin pairings have you got?

[javadjamshidi]

Yes, but I couldn't find a clear answer, most of them excluded the relatives though. It's not a big sample ( ~1500) but generally, I want to know if using a GRM in a mixed model can correct the effect of such relatedness or not.

[Kevin Blighe]

There are contradictions everywhere in this field, which is why there is never a clear answer. I read another twins study where no correction for family was made. If you feel that you need to account for the family-specific variation then proceed with the GRM.

[javadjamshidi]

Thanks Kevin, yes, it seems there is not a clear answer.