|
|
Post by hugonob on Sept 15, 2015 22:06:12 GMT
Hi,
I would like to know if it is correct transform effect size from beta to OR using exp(beta) = OR. I did a case-control association study with --mlma
Thanks
Hugo
|
|
|
|
Post by Jian Yang on Sept 17, 2015 0:54:43 GMT
There is a transformation equation in the Supplementary Note of
Cortes et al. (2013) Identification of multiple risk variants for ankylosing spondylitis through high-density genotyping of immune-related loci. Nat Genet, 45:730-738.
OR=1+b01*p(1-p)/[v(1-v)p2(1-p1)].
I derived this equation. It's not optimal because it uses allele frequencies in cases and controls. It will be great if someone could come up with a better solution.
|
|
|
|
Post by natvoc on Jun 17, 2021 14:21:13 GMT
Hi! Am writing to ask of (I think) a similar question:
If I simulated the betas of some SNPs, is it proper to convert the effect sizes into ORs with @jianyang's formula, OR=1+b01*p(1-p)/[v(1-v)p2(1-p1)]
And if it is, I would assume b01 is the beta of said SNP, p would be the allele frequency, but what are v, p1 and p2?
|
|
