A Wavelet-Based Nonparametric Approach to Association Analysis of Functional Data

• Activity Number: 512
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Heejung Shim*+ and Matthew Stephens
• Companies: The University of Chicago and The University of Chicago
• Keywords: RNA-Seq ; functional data ; wavelet regression ; gene expression ; association analysis ; Bayesian nonparametric approach

Abstract

High-throughput sequencing technologies are now routinely applied at a genome-wide scale to collect a variety of phenotypic data. For example, RNA-Seq and DNase-Seq are used to assay overall gene expression and chromatin accessibility, respectively. Here we consider the development of statistical methods for testing for differences in these types of "functional" data between two or more groups (e.g. testing for genetic variants that affect overall gene expression). In this work we introduce a Bayesian nonparametric approach to this problem, based on a wavelet regression known to be well suited for modeling spatially heterogenous functional data. In addition to testing for association, our approach aims to provide a better interpretation of the analysis such as which parts and features of the functional data are associated with a given variant. Unlike alternative Bayesian approaches that use MCMC sampling, our method allows Bayesian inference to be performed analytically, which makes application to genome-wide studies practical. We illustrate the proposed method on RNA-Seq data from 69 LCLs derived from Nigerian individuals generated as part of the International HapMap project.