Nonparametric Covariate-Adjusted Association Tests Based on the Generalized Kendall's Tau()

Area of Science:

• Psychiatric research
• Genetic epidemiology
• Statistical genetics

Background:

• Comorbidity risk factor identification is crucial in psychiatric research.
• Simultaneous analysis of multiple, correlated traits offers greater power than single-trait analysis.
• Complex diseases, particularly mental illnesses, involve traits measured on various scales (dichotomous, ordinal, quantitative).

Purpose of the Study:

• To propose a novel nonparametric test for multiple complex traits that adjusts for covariate effects.
• To develop a method that optimally adjusts for covariates in association studies involving traits of different scales.
• To provide a robust statistical tool for analyzing comorbidity in genetic studies.

Main Methods:

• Developed a nonparametric association test for multiple complex traits incorporating covariate adjustment.
• Utilized a maximum statistic derived from multiple adjusted test statistics for optimal adjustment.
• Derived the asymptotic null distribution and proposed a resampling approach for significance assessment.

Main Results:

• The proposed nonparametric adjusted test demonstrated increased power in the presence of environmental effects.
• The test showed greater robustness to model misspecifications compared to existing parametric adjusted tests.
• Simulations confirmed the type I error and power advantages of the new test.

Conclusions:

• The developed nonparametric test offers a powerful and robust method for analyzing comorbidity in genetic studies with covariates.
• This approach is particularly valuable for complex traits measured on different scales.
• The method was successfully applied to a dataset on the genetics of alcoholism, highlighting its practical utility.