|
Abstract:
|
In recent genetic association studies, statistical methods to identify pleiotropic variants associated with multiple phenotypic traits have been developed, since susceptible variants with small effects could be easily missed in association studies based on a single trait. However, most of the existing methods to identify pleiotropic variants are designed for only quantitative traits even though pleiotropic variants are often associated with both quantitative and qualitative traits. There are some meta-analysis methods which basically integrate summary statistics of individual variants associated with either a quantitative or qualitative trait. But, these methods cannot account for correlations between genetic variants. In this article, we propose new selection probability computation to prioritize individual variants associated with both quantitative and qualitative traits. For each phenotypic trait, coefficients of elastic-net regularization are first estimated and then they are additively combined to compute selection probability of individual variants. We demonstrated that the proposed methods outperform the existing methods in both simulation studies and real data analysis.
|
