Map PRS score to EID
Hi all,
After generating the PRS from the imputation data, how can we link the IID from plink output PRS file (which are -1, -2 ) to the EID in the phenotype data (EID/Patient ID) so we can run association test? (Is gfetch to the local machine the only way to get the actual .sample file? )
Thanks for your help!!
Comments
2 comments
I would expect to see the first few hundred rows of output with -1, -2 etc. These correspond to participants who have withdrawn from the UKB. I would expect the other rows to have EID in the IID column. If this is not what you see, please explain a bit more about what plink command you ran.
Hi,
Thanks for your advice, I checked the plink output and the original .sample file downloaded from the UKB, the IID are from -1 to -487409… The below is how I generated the PRS:
1. convert the .bgen to plink2 trio files
plink2 --bgen UKB_imputation_all_chr.bgen ref-first --sample ukb22828_c1_b0_v3.sample --set-all-var-ids @:#\$r\$a --freq --make-pgen --out raw
2. generate PRS
plink2 --pfile raw --score PRScsx_SCZ_META_all_chr_withID.txt 7 4 6 cols=scoresums no-mean-imputation -out Fullset_PRS
Thanks again for your help!
Please sign in to leave a comment.