DSuveges
commented
2 months ago Essentially we need to do LD annotation for the lead variants, as they were coming from a study with 100% Europeans instead of 70%. There's already a method to get the major ancestry from the ld popultation field in the study index. The steps that need to be touched apply the ld_annotate method.
As previously discussed several times the logic/math behind combining multiple ancestries in PICS fine-mapping is incorrect. PICS will benefit from using the same approach as used now in susie - use only major ancestry. In the future we can think about refining the logic, but for now it is better to be as correct as possible.