Bayesian Functional Data Methods in Copy Number Alteration Studies: Applications in Urothelial Bladder Carcinoma

• Activity Number: 429
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Miranda Lynch* and Jessica M. Clement
• Companies: University of Connecticut Health Center and University of Connecticut Neag Comprehensive Cancer Center
• Keywords: functional data regression ; bladder cancer ; copy number alterations ; Bayesian inference

Abstract

Chromosomal level alterations in the genome are a hallmark of cancer, and methods to probe copy number alterations (CNAs) have revealed a number of important drivers of oncogenesis and tumor progression. This work is motivated by our research questions in urothelial bladder carcinoma (BLCA), investigating the interconnections between smoking status/history and immune response in BLCA progression. In this work, copy number profiles are characterized using Bayesian functional data methods employing wavelet basis functions. These basis functions are well suited for the types of profiles that appear in copy number studies using array CGH and SNP arrays. We propose methods using these profiles for functional regression to examine the relationship to smoking and to look at whether the altered genomic regions preferentially include genes indicative of immune response. We apply our methods to publicly available bladder cancer data from a group of patients with metastatic disease.