Proposal - serious example

[elecarnero]

I would propose to retain the GWAS results of genetically-determined methylation PCs, as they could contain interesting information of genetic variants (trans) influencing global methylation. We discussed also this briefly at ECD.

[explodecomputer]

I think this is a really good idea, and certainly straightforward to implement. The only issue that we would have to think about is how to do the meta analysis because the genetically determined PCs generated in one study won't necessarily capture the same things as in another study, it assumes:

• all genetically determined methylation axes are the same rank in each study
• the power to detect a genetic PC is the same in different studies

Is there something we can do here with the loadings? Like, we compare the loadings of each PC between each cohort and decide which ones to meta analyse together that way?