Analysis of Genetic Association Studies Incorporating Prior Information of Genetic Models

Abstract

In genetic association studies, for each underlying genetic model, there is an optimal test. Usually, the true model is unknown, but knowledge from previous genome-wide association studies for the disease under investigation is available and provides information about the underlying model. We study how to incorporate this prior information about plausible genetic models to achieve better efficiency robustness in genetic association studies. Two procedures are proposed and studied. The first one reduces the set of genetic models using the prior information. The second one derives posterior probabilities for the plausible genetic models using trend tests. Then the trend test based on the posterior mean of the genetic model or a weighted trend test over various genetic models can be employed, similar in spirit to the efficiency robustness approach. In the proposed procedures, the strong Hardy–Weinberg disequilibrium in cases is studied. Simulations are conducted to compare the proposed methods with existing ones. The usefulness of the results for the analysis of data collected in replication studies is investigated, the proposed methods are shown to have better overall performance than existing one, and the methods are applied to analyze the real data.