search

omerwe / polyfun

PolyFun (POLYgenic FUNctionally-informed fine-mapping)
MIT License
96 stars 22 forks source link

Alternative to HESS estimate #123

Closed jerome-f closed 2 years ago

jerome-f commented 2 years ago

Hi Omer,

I have a recommendation to substitute for HESS generated priors in the SuSie workflow. We discussed on other different threads as to how this causes issue with small sample sizes. One recommendation I have is to take the heritability form the FINEMAP output, FINEMAP is quite robust in estimating regional heritability even in small N cases. Of course this is moot if h^2 is calculated with S-LDSC. But if we want to use the h^2 estimate with out functional finemapping this would be a better alternative.

omerwe commented 2 years ago

Hi @jerome-f, This sounds like a great idea. I never tried using the FINEMAP estimates with SuSiE, but I don't see a reason why it wouldn't work... But I guess a bigger question is --- if you run FINEMAP anyway, do you have a good reason to run SuSiE afterwards?

In any case, I'm afraid I don't have the bandwidth to implement this, but I'm happy for someone to step in and submit a PR.