Multi-Cohort, Network-Guided Regression Approach for Genetic Interaction Detection

• Activity Number: 663
• Type: Invited
• Date/Time: Thursday, August 8, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Josee Dupuis*+ and Chen Lu and Eric Kolaczyk
• Companies: Boston University School of Public Health and Boston University School of Public Health and Boston University
• Keywords: genetic association ; gene-gene interaction ; network ; sparse regression

Abstract

Despite the success of genome-wide association studies in identifying genetic variants associated with complex traits, most often only a small portion of the total heritability is explained by the associated loci. Because complex diseases are believed to be caused by a number of genetic and environmental factors, progress in fully characterizing the genetic architecture of complex traits may come from unraveling the interplay of genes with each other, as well as with environment, through the system of biological pathways and related networks. We present a novel, network-guided statistical methodology to facilitate the discovery of gene-gene interactions associated with complex quantitative traits. Our method uses sparse regression principles to fit regression models (i.e., of phenotype on genetic markers) with second-order interactions, based on a penalized least-squares criterion, where the penalty incorporates known network biology in the form of pathways and gene function. We present an extension of the applicability of this regression framework to multiple cohorts through the development of a two-stage meta-analysis strategy. We assess the overall methodology in simulation.