Cancer Marker Identification via Penalized Integrative Analysis

• Activity Number: 189
• Type: Invited
• Date/Time: Monday, August 2, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Shuangge Ma*+
• Companies: Yale University
• Keywords: integrative analysis ; cancer genomics ; marker selection

Abstract

In cancer genomic studies, marker identification from analysis of a single data source has suffered from lack of reliability and reproducibility. More and more genomic studies are being shifted towards data integration and integrative analysis. Most existing methods are ad hoc, and there is a lack of novel methods with sound statistical basis. In this study, we propose novel penalization methods for integrative analysis of cancer genomic data from multiple sources and marker identification. The proposed methods can automatically accommodate the heterogeneity across multiple sources and the joint effects of multiple markers. We rigorously establish asymptotic selection properties and conduct simulations. Analysis of data on multiple cancers shows satisfactory performance of the proposed methods.