RNOmni: Rank Normal Transformation Omnibus Test

Implementation of genetic association tests that incorporate the rank-based inverse normal transformation (INT). These tests are broadly indicated for traits with continuous residual distributions. In the presence of non-normal residuals, INT-based tests robustly control the type I error, whereas standard linear regression may not. Moreover, INT-based tests dominate standard linear regression in terms of power. There are two main strategies for incorporating the INT in association analysis. In direct INT (D-INT), the trait is directly transformed. In indirect INT (I-INT), residuals are formed prior to transformation. Neither D-INT nor I-INT is uniformly most powerful. The INT omnibus test (O-INT) adaptively combines D-INT and I-INT into a single robust and statistically powerful approach.

Downloads:

Linking:

Please use the canonical form https://CRAN.R-project.org/package=RNOmni to link to this page.