Latent Variable Models for Assessing Individual Drug Sensitivity and for Predicting Hereditary Disease Risk

• Activity Number: 149
• Type: Invited
• Date/Time: Monday, July 30, 2012 : 10:30 AM to 12:20 PM
• Sponsor: WNAR
• Author(s): William Evan Johnson*+ and Stephen Piccolo and Andrea Bild
• Companies: Boston University and University of Utah and University of Utah
• Keywords: personalized medicine ; mixture models ; factor analysis ; meta-analysis ; gene expression ; Bayesian analysis

Abstract

The development of personalized treatment regimes is an active area of current research in genomics. The focus of our research is to investigate core biological components that contribute to disease prognosis and development, and to develop latent variable models to accurately determine optimal therapeutic regimens for individual patients. In this research we have developed a novel gene expression 'barcoding' approach, which classifies genes as active/inactive, and improves researchers ability to characterize pathways, cluster patients and link prognosis to pathway status in individual patients. Our barcoding method creates a probabilistic barcode that is applicable to all microarray and sequencing platforms, allowing for efficient combination of data from multiple labs, experiments, or platforms and facilitating downstream analysis. We are currently working on a variety of applications using data from high-risk breast cancer cohorts, families with high risk of psychiatric disorders, and a pedigree with a rare X-linked genetic disease.