An Empirical Bayes Approach for Joint eQTL Analysis in Multiple Tissues

• Activity Number: 494
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 8:30 AM to 10:20 AM
• Sponsor: IMS
• Author(s): Gen Li*+ and Andrey Shabalin and Ivan Rusyn and Fred Wright and Andrew Nobel
• Companies: UNC-CH and Virginia Commonwealth University and UNC-CH School of Public Health and The University of North Carolina and UNC-CH
• Keywords: Empirical Bayes ; Hierarchical Model ; eQTL Analysis ; Tissue Specificity ; Local FDR

Abstract

The analysis of expression quantitative trait loci (eQTL) aims to identify genetic variants that regulate gene transcription, and potentially help dissect complex transcriptional-based mechanisms of disease. Patterns of transcriptional variation can be highly tissue-specific, and identifying tissue-common and tissue-specific eQTLs is of great interest. However, most existing methodologies for multiple-tissue eQTL analysis have been informal and post-hoc. We propose an empirical Bayes approach for joint eQTL analysis in multiple tissues that is based on a hierarchical model for the observed correlations of gene-SNP pairs across the available tissues. Once fit, the model provides interpretable posteriors for eQTLs across the range of tissues in the original data set. Via a data-based simulation, we demonstrate that the joint analysis yields substantial gains in statistical power within individual tissues and improves assessments of specificity across tissues when compared with post-hoc tissue-by-tissue analysis. We also present results from the analysis of real data.