Robust Estimation for Secondary Trait Association in Case-Control Genetic Studies

• Activity Number: 139
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Jean de Dieu Tapsoba*+ and Charles Kooperberg and Alexander P. Reiner and Ching-Yun Wang and James Y. Dai
• Companies: Fred Hutchinson Cancer Research Center and Fred Hutchinson Cancer Research Center and University of Washington and Fred Hutchinson Cancer Research Center and Fred Hutchinson Cancer Research Center
• Keywords: Case-control sampling ; Inverse probability weighting ; Maximum likelihood

Abstract

Case-control genetic association studies are often nested in large cohort studies, which also collect extensive data on secondary traits e.g., biomarkers and quantitative phenotypes. Secondary trait association adds insight to the genetic architecture of disease etiology, but requires caution in estimation. Ignoring case-control sampling may introduce bias to the estimation of the secondary trait association. We compare maximum likelihood (ML) and inverse probability weighted (IPW) methods, which are commonly used in secondary trait analyses. We show that a misspecified primary trait model using the ML method can severely inflate the type I error, while IPW based estimators are always consistent and yield proper control of false positive rates. Also, the IPW using nonparametrically estimated selection probabilities and the augmented IPW improve efficiency over the simple IPW when the secondary trait is available for the entire cohort. We conclude that in large genetic association studies with complex sampling schemes, IPW based estimators offer flexibility and robustness, and therefore are a viable option in analysis.