A fully adjusted two-stage procedure for rank-normalization in genetic association studies: SOFER et al.

When testing genotype–phenotype associations using linear regression, departure of the trait distribution from normality can impact both Type I error rate control and statistical power, with worse ...