Analysis of High-Throughput Methylation Data Using Wavelet-Based Functional Mixed Models

• Activity Number: 393
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Wonyul Lee*+ and Jeffrey S. Morris
• Companies: MD Anderson Cancer Center and MD Anderson Cancer Center
• Keywords: Functional mixed models ; Wavelet transform ; Methylation ; Bayesian analysis

Abstract

DNA methylation is a key regulator of gene function in biological process. Deciphering the DNA methylation code has recently received much attention while many of works have focused primarily on the individual gene level. Due to recent advance in high-throughput platform for DNA methylation assessment, comprehensive genome-wide picture of DNA methylation can be investigated. A common approach to perform genome-scale analysis of DNA methylation data is to model independently each genomic location, which fails to appropriately capture the special correlations across genome. In this work, two high-throughput methylation data generated by CHARM microarrays are analyzed using the wavelet-based functional mixed model introduced by Morris and Carroll (2006) to properly capture the spatial correlations across genome. We demonstrate that the wavelet-based mixed model has better performance than the independent approach as it properly takes the correlation into account.