Identifying Gene Regulatory Networks Using Multiple-Tissue EQTL Data

• Activity Number: 495
• Type: Topic Contributed
• Date/Time: Wednesday, August 1, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Jichun Xie*+ and Tony Cai and Hongzhe Li and Weidong Liu
• Companies: University of Pennsylvania School of Medicine and University of Pennsylvania and University of Pennsylvania and Shanghai Jiao Tong University
• Keywords: conditional dependency ; eQTL data ; multiple tissue

Abstract

Precision matrices reveal conditional dependency structures between multivariate normal random variables. Some heterogenous data share similar dependency structures among different groups. One of the examples is the multiple-tissue gene expression data. Evidence has shown that the gene regulatory networks share similarities across different tissues and yet have tissue-specific interaction mechanism. To improve the estimation efficiency, we de velop a joint estimation method to infer precision matrices. Theoretical and Numerical studies show that our method has improved estimation efficiency by the scale $\sqrt{log(K)/K}$, where $K$ is the number of the tissues. Our method also has better support recovery results compared with seperate estimation methods as well as other competing joint estimation methods. We further apply our method to the mouse gene expression data and obtain the gene regulatory networks for multiple mouse tissues.