Bayesian Hierarchical Models for Data Integration in Genomics

• Activity Number: 383
• Type: Topic Contributed
• Date/Time: Tuesday, July 31, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Franceso Stingo*+
• Companies: MD Anderson Cancer Center
• Keywords: Bayesian hierarchical models ; High-dimensional data ; Genomics ; Graphical models ; Bayesian variable selection ; Data integration

Abstract

In this talk I will focus on recent prior constructions that can be used for the analysis of cancer related high-throughput data. I will describe Bayesian approaches that allow for the integration of different sources of biological information into the analysis of experimental data. I will then explain how a graphical model approach can account for the biological relations among the variables of interest even when the number of observed variables is much greater than the sample size. This approach can be successfully applied to several types of biological networks like gene networks, networks based on the amount of linkage disequilibrium (LD), and regulatory networks. The value of these methods is that they can provide new insights into the analysis of experimental data, test new types of hypotheses, take into account complex dependence structures among several variables, and in general, provide a better understanding of the biological processes involved in the development of cancer.