search

broadinstitute / tensorqtl

Ultrafast GPU-enabled QTL mapper
BSD 3-Clause "New" or "Revised" License
162 stars 52 forks source link

Allele frequency reported in nominal cis-mQTLs output #88

Closed ariadnacilleros closed 1 year ago

ariadnacilleros commented 1 year ago

Hi,

I have run two analyses with the nominal cis-QTL approach. In both analyses, I used the same input files, except for the covariates file, which was different. But in the outputs from the analyses, different AF has been reported. In one of them, all the AF correspond to the MAF, and in the other, not always. So, I was wondering, if the effect allele considered is the one corresponding to the AF column, or if it is the ALT from the bim file (PLINK)? And also, how this could happen if both analyses have been run with the same PLINK set files? Thank you so much for the tool.

Best regards,

Ariadna.

francois-a commented 1 year ago

Hi, sorry for not responding to this — the effect is always the ALT allele. I you see discrepancies it's likely due to an allele flip in PLINK 1.9 when the --keep-allele-order flag is not used.