What is the effect size in output of FaST-LMM when using EWASher?

Question

[s.w.vanderlaan]

Hi,

Perhaps a naive question, but it's just not clear to me. From the adapted FaST-LMM program EWASher I get this output.

SNP Chromosome GeneticDistance Position    Pvalue    Qvalue   N NullLogLike AltLogLike SnpWeight   SnpWeightSE OddsRatio WaldStat NullBias NullCov00Weight NullCov01Weight NullCov02Weight NullCov03Weight   NullCov04Weight NullCov05Weight NullCov06Weight NullCov07Weight NullCov08Weight NullCov09Weight   NullCov10Weight NullCov11Weight NullCov12Weight NullVar
rs12345 5 395444   395444 1.216e-32 4.658e-27 475 -310.2  -239   -0.2938     0.02281     2.272 166   0.4021   -0.01345   -0.06752    -0.0474  0.0003574  0.0002916   0.000799 -0.0002633   0.001334  0.0003579  0.0005426  0.0003502  -0.001946   0.000124  0.2161

What are the effect size (beta), and corresponding standard error? Is that SnpWeight, or OddsRatio? If I calculated exp(-0.2938), I do not get OddsRatio=2.272, conversely if I do ln(2.272) it's not immediately clear to which column this corresponds? The model I used contained a couple of covariates, perhaps the OddsRatio is that of the full model, and SnpWeight is that of the model without covariates, or vice versa?

It's not immediately clear to from the paper either.

Thanks and best,

Sander

[mbrozynska]

Hi @s.w.vanderlaan, have yo found the answer to your question?