A Latent Variable Model for Integrative Clustering Analysis of Multi-Type Genomic Data

• Activity Number: 16
• Type: Topic Contributed
• Date/Time: Sunday, August 3, 2014 : 2:00 PM to 3:50 PM
• Sponsor: International Society for Bayesian Analysis (ISBA)
• Author(s): Qianxing Mo*+ and Ronglai Shen and Sijian Wang and Venkatraman Seshan and Adam Olshen
• Companies: Baylor College of Medicine and Memorial Sloan Kettering Cancer Center and University of Wisconsin and Memorial Sloan Kettering Cancer Center and University of California, San Francisco
• Keywords: Integrative clustering analysis ; Bayesian methods ; Genomic data ; Cancer data ; High dimensional data ; TCGA

Abstract

Large-scale genomic profiling studies such as The Cancer Genome Atlas (TCGA) project have generated multi-type genomic data (e.g., gene expression, methylation, aCGH, sequencing data, etc.) for individual tumor samples. In an effort to identify clinically relevant tumor subtypes, we developed a latent variable model that can jointly model four different data types including binary, categorical, count and continuous data to cluster tumor samples based on their joint profiling patterns. The method combines Bayesian and frequentist approaches in order to achieve genomic feature selection and joint dimension reduction of complex genomic data to a few eigen features that can be used for integrated visualization and cluster discovery. We will use the TCGA and the cancer cell line encyclopedia data to illustrate the method.