A Bayesian Mixture Model for the Integration of Multiple High-Throughput Genomic Platforms

• Activity Number: 657
• Type: Contributed
• Date/Time: Thursday, August 4, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Filippo Trentini*+ and Yuan Ji and Peter Mueller
• Companies: Bocconi University and The University of Texas MD Anderson Cancer Center and The University of Texas MD Anderson Cancer Center
• Keywords: gene expression ; aCGH ; bayesian mixture models ; latent factors ; MCMC

Abstract

We consider modeling jointly gene expression microarray and array CGH data. We propose a Bayesian mixture model for both the observed copy number and gene expression measurement that uses latent categories, and the integration between the two platforms is done via a regression of the latent Gaussian probit scores for gene expression microarray data on latent Gaussian probit scores for aCGH data. In the regression covariates such as biological conditions are introduced to make inference on gene differential expression based on the integration of copy numbers and mRNA abundances. This models also generates natural similarity measures for traditional clustering, and gives probabilistic statements about the assignment of tumors to molecular profiles. It has been applied to a novel data set consisting of 122 newly diagnosed breast cancer patients.