If one wanted to use the five psychiatric latent factors from the Nature paper for PRS analyses, and the cohort of interest contributed to the discovery GWASs used to derive these factors, would it be possible to obtain factor GWAS summary statistics that exclude this cohort, or otherwise account for sample overlap when calculating PRSs?
Replying to “If one wanted to use the five psychiatric latent f…”:
If you are working within that cohort your cohort will usually have been given a leave one out (LOO) or you can ask for one