Discovering Influential Variables: A Partition-Based Learning Method to Identify Susceptible Genetic Risk Factors in Common Human Disorders

• Activity Number: 271
• Type: Invited
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: ENAR
• Author(s): Tian Zheng*+ and Herman Chernoff and Shaw-Hwa Lo
• Companies: Columbia University and Harvard University and Columbia University
• Keywords: measure of influence ; high dimensional data ; statistical genetics ; breast cancer

Abstract

Due to advances in technology, there has been an increasing availability of high dimensional genetic data. One challenge is to develop effective methods of finding useful information from the vast amounts of messy and noisy data. In this talk, we discuss a general computer intensive approach for detecting which, of many potential explanatory variables, have an influence on a dependent variable Y. This approach is suited to detect influential variables, where causal effects depend on the confluence of values of several variables. It has the advantage of avoiding a difficult direct analysis, involving possibly thousands of variables, by dealing with many randomly selected small subsets from which smaller subsets are selected, guided by a measure of influence. Applications of this general approach to a epidemiology study and a gene expression study of breast cancer will also be discussed.