I am writing to double check the reference allele used when providing the effect size in rvtest summary statistics.
The description in the manual reads: "ALT_EFFSIZE: for continuous outcome, this is the estimated effect size; for binary outcome, this is the estimated log odds-ratio. We apply a new correction method when binary trait associations for related individuals are analyzed in standard linear mixed models. The log odds ratio is approximately correct for related individual analysis as well."
Does this mean that in a binary trait association, the value I get in the ALT_EFFSIZE column is the log odds-ratio when having the ALT ALLELE with respect to the REF ALLELE? Or is is the opposite? I am building a PRS using my results from rvtest and the direction seems to be opposite to what I would expect.
Hello,
I am writing to double check the reference allele used when providing the effect size in rvtest summary statistics.
The description in the manual reads: "ALT_EFFSIZE: for continuous outcome, this is the estimated effect size; for binary outcome, this is the estimated log odds-ratio. We apply a new correction method when binary trait associations for related individuals are analyzed in standard linear mixed models. The log odds ratio is approximately correct for related individual analysis as well."
Does this mean that in a binary trait association, the value I get in the ALT_EFFSIZE column is the log odds-ratio when having the ALT ALLELE with respect to the REF ALLELE? Or is is the opposite? I am building a PRS using my results from rvtest and the direction seems to be opposite to what I would expect.
Thank you for your time in advance.