Abstract:
|
Genome-wide Association Studies (GWASs) for complex diseases often collect data on multiple correlated traits. Multivariate analysis of these traits can improve power to detect genetic variants associated with the underlying disease. Multivariate analysis of variance (MANOVA) is a popular approach but the behavior of MANOVA under different trait models has not been carefully investigated. We showed, both theoretically and using simulations, that MANOVA is generally very powerful for detecting association but there are situations where MANOVA may completely lose power. We propose a unified score-based test USAT that can perform better than MANOVA in such situations and nearly as well as MANOVA elsewhere. USAT reports an approximate asymptotic p-value for association and is computationally efficient to implement at a GWAS level. We have extensively studied the performance of USAT and other existing methods, and demonstrated the advantage of using USAT to detect association between a genetic variant and multivariate traits. We applied USAT to data from three correlated traits collected on 5,816 Caucasian individuals from the ARIC Study and detected some interesting association.
|