DNA methylation is an important genomic modification involved in the regulation of gene expression, cell differentiation and developmental processes. Abnormal methylation has been implicated in human diseases such as various types of cancers. There has been great interest in identifying genes with differentially methylated regions (DMRs) between different biological conditions. With the high-throughput technologies, large-volumes of methylation data at genomic level have been generated. This also poses challenges in data management, analysis, and interpretation. Taking advantage of the high-quality genomic data, various strategies are employed for detecting functional differentially methylated regions (fDMRs) by integrating both methylation data and gene expression data. One such approach is based on unified Bayesian hierarchical modeling. We model DNA methylation and gene expression in a hierarchical way, based on the assumption that fDMRs will show changes in methylation levels, which then lead to changes in gene expression. The evidence of differential methylation will be used as prior information to update the evidence of differential gene expression. Integrating methylation and gene expression data will help address some of the challenges in DMR identification methods and more effectively identify fDMRs, which can help elucidate the genetic mechanisms of complex human diseases.