I have downloaded a PRS from the PRS catalogue that has been published and validated in a cohort from the UK Biobank.
I would like to run this on a different cohort in the UK 100,000 genomes project.
Many of the online PRS guides focus on creating and QCing the PRS which I do not need to do.
I want to compare 400 cases with 20000 controls, they are stored in a phenotype file with their top ten PCs as covariates alongside sex. Their genotype data is stored in multi-sample VCFs and PGEN files chunked by position (there are 1500 chunks crossing the whole genome).
Looking at LDPred it seems hard to "skip" the first part and just provide the score and the files to get the PRS for each individual. Is that the case? Would a different tool be better? How would this code look being run on a per chunk basis.
Many thanks your help!