search

omerwe / polyfun

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

Per Locus #12

Closed bschilder closed 4 years ago

bschilder commented 4 years ago

Kind of a random (and possibly dumb) question; does it make any sense to run LDSC on a single user-defined locus (e.g. a locus from a GWAS, that spans anywhere between 500 and 5000 bps)? Or does LDSC always require genome-wide data?

omerwe commented 4 years ago

Good question Unfortunately LDSC requires genome-wide data. For a single locus it will return extremely noisy estimates (basically just noise), and the standard error estimates will also be wrong...

bschilder commented 4 years ago

I see, that makes sense. Thanks!