Graph-Based Interaction Association Mapping in Genome-Wide Studies

• Activity Number: 57
• Type: Invited
• Date/Time: Sunday, July 31, 2011 : 4:00 PM to 5:50 PM
• Sponsor: National Institute on Drug Abuse-NIH
• Author(s): Yu Zhang*+
• Companies: Penn State University
• Keywords: Association Mapping ; Epistasis ; Graph model ; Bayesian method

Abstract

Genomewide association studies are becoming increasingly important due to the advance in high-throughput sequencing technologies. In addition to detecting marginal associations, it is also of interests to identify multi-marker interaction associations. Mapping from an astronomical number of possible interactions in the genome is a daunting task both computationally and statistically. For high-density markers, the problem is further complicated by the complex dependence between markers. We introduce a graph-based Bayesian model for large-scale interaction association mapping. Compared with existing methods, our method has two features. 1) We identify complex gene-gene interaction graphs associated with the disease. 2) We design better models to account for the dependence in high-density markers. We use simulation and real data examples to demonstrate the performance of our method.