Abstract:
|
In genetic association testing, failure to properly adjust for population structure can lead to severely inflated type I error and loss of power. Meanwhile, adjustment for relevant covariates is often desirable and sometimes necessary to protect against spurious association and to improve power. Many recent methods to account for population structure and covariates are based on linear mixed models (LMM), primarily designed for quantitative traits. For binary traits, however, LMM is often a mis-specified model and can lead to power loss. We develop a new method for binary trait association testing using a quasi-likelihood framework, which exploits the dichotomous nature of the trait by modelling covariate effects on a logit scale and achieves computationally efficiency through estimation equations. P-values are assessed retrospectively to ensure robust adjustment for population structure. We show through simulation studies that our method provides power improvement over the linear mixed method approach in a variety of population structure settings and trait models. The method is illustrated in an association analysis for Crohn disease in the WTCCC dataset.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.