|
Abstract:
|
Empirical findings suggest many trait-influencing genetic factors are associated with multiple distinct phenotypes; these associations are defined as pleiotropic or cross-phenotype associations. When cross-phenotype effects exist, association methods of multiple phenotypes that leverage pleiotropy are more powerful than univariate methods that consider each phenotype separately. We recently developed a method called the Gene Association with Multiple Traits (GAMuT) test, which is a nonparametric distance-covariance approach for statistical inference. The method is scalable to genome-wide analysis, can accommodate an arbitrary number of phenotypes, and also can adjust for influential covariates. In my talk, I first provide a brief introduction to GAMuT and then discuss how to modify the framework to perform mediation analysis. I will also discuss GAMuT extensions to other study designs as well as extensions to handle complex questionnaire data collected in psychiatric studies
|
