Learning Regulatory Network from Multiple Genomics Array Data

• 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): Pei Wang*+ and Jie Peng
• Companies: Fred Hutchinson Cancer Research Center and University of California, Davis
• Keywords: network ; directed graph ; genomics array ; sparse regression

Abstract

When studying genomic changes in disease samples, it is of great interest to distinguish primary events driving diseases from secondary events resulting from diseases. We propose to tackle this problem by building a directed graph for characterizing the relationships among DNA and RNA activities measured for thousands of genes. With proper assumptions, we show that the directions of interactions can be revealed by examining the dependence strengths in a joint regression model. We also allow the model to incorporate prior biological knowledge of the interaction directions. We illustrate the performance of the method through simulation studies and one real application.