Joint Inference of Sparse Network and Genetic Association in Genetical Genomics Studies

• Activity Number: 145
• Type: Invited
• Date/Time: Monday, August 2, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Hongzhe Li*+
• Companies: University of Pennsylvania
• Keywords: high dimension ; network ; regularization ; eQTL ; sparse seemingly unrelated regression ; Gaussian graphical models

Abstract

Genetical genomics experiments have been routinely conducted to measure both the genetic variants and gene expression data on the same subjects. The gene expression levels are often treated as quantitative traits and are subject to standard genetic analysis. However, the genetic architectures for many gene expressions may be complex, and poorly estimated genetic architectures may compromise the inferences of the dependency structures of the genes at the transcriptional levels. We introduce a joint modeling approach in the framework of sparse seemingly unrelated regression models to simultaneously identify the genetic variants associated with gene expressions and construct the sparse Gaussian graphical model based on the eQTL data. We present an efficient coordinate descent algorithm, simulation studies and application to analysis of a yeast eQTL data set and theoretical results.