Mixed-Effects Logistic Model for Association Analysis

• Activity Number: 536
• Type: Contributed
• Date/Time: Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Sanjay Shete*+
• Companies: The University of Texas M.D. Anderson Cancer Center
• Keywords: Mixed-Effects model ; case-control study ; association ; covariates ; power

Abstract

An association study to identify possible causal genes is a popular approach. However, in association studies cases and controls are assumed to be genetically unrelated. In certain situations, genetic data may be available on other affected family members. Choosing a single affected individual per family is statistically inefficient and leads to a loss of power. On the other hand using affected family members and unrelated normal controls directly leads to false-positive results. We propose a new approach using mixed-model logistic regression, in which associations are performed using family members and unrelated controls. Extensive simulation studies showed that our approach can effectively control the type-I error probability and has higher power than likelihood based methods. We applied this method to correlate mutagen sensitivity and genes involved in NER pathway using a twin study.