Two independent phenotypes simulated under the null hypothesis of no 
genotypic effect. A subject specific linear predictor was calculated based on
age, sex, pc1, and pc2. The normal phenotype was generated by adding N(0,1)
deviates to the linear predictor, while the T3 phenotype was generated by
adding T3 deviates to the linear predictor.

datasets

Implementation of genetic association tests that utilize the rank based inverse normal transformation (INT). The primary contribution is an omnibus test, which synthesizes two INT-based approaches, termed direct INT (DINT) and partially indirect INT (PIINT). In genome scale simulations, the omnibus test routinely provided valid inference against continuous phenotypes with skewed and heavy tailed residual distributions. Association testing against these same phenotypes via standard linear regression led to an excess of false positive associations. In addition to controlling the type I error, the omnibus approach provided power comparable to the more powerful of the component methods.

Zachary McCaw

RNOmni

Omnibus Test for Genetic Association Analysis using the Rank
Normal Transformation

Y function

A numeric matrix with 1000 rows and 2 columns<dl class="dl-horizontal">
 <dt>YN</dt><dd>Normal phenotype.</dd>
 <dt>YT3</dt><dd>T3 phenotype.</dd>
</dl>

Format

Simulated Phenotypes — Y

A numeric matrix with 1000 rows and 2 columns<dl class='dl-horizontal'>
 <dt>YN</dt><dd>Normal phenotype.</dd>
 <dt>YT3</dt><dd>T3 phenotype.</dd>
</dl>

Y: Simulated Phenotypes

Description

Usage

Arguments

Format