Bayesian Spatial Functional Models for High-Dimensional Genomics Data

• Activity Number: 203
• Type: Invited
• Date/Time: Monday, August 4, 2014 : 2:00 PM to 3:50 PM
• Sponsor: SSC
• Author(s): Veera Baladandayuthapani*+ and Jeffrey S. Morris and Lin Zhang and Keith Baggerly
• Companies: MD Anderson Cancer Center and MD Anderson Cancer Center and MD Anderson Cancer Center and MD Anderson Cancer Center
• Keywords: bayesian ; functional ; wavelets ; genomics ; MCMC ; spatial

Abstract

Many scientific applications generate correlated functional data, where it is often of interest to flexibility model the dependence patterns. We present methods that focus on spatial functional data analysis where spatial dependence is present in high-dimensional functional data. Our methods allow for simultaneous characterization of the high-dimensional functions using non-parametric basis functions, joint modeling of spatially correlated functional data and detection of local features in spatially heterogeneous functional data - to answer several important biological questions. Our methods are motivated by and applied to a high-throughput copy number dataset generated through whole-organ histologic genomics maps of bladder cancer development. Our model identifies several genetic markers with copy number alterations that are potentially associated with development of bladder cancer, and not discovered by methods that ignore the spatial dependence.