mCSEA: detecting subtle differentially methylated regions

Bioinformatics. 2019 Sep 15;35(18):3257-3262. doi: 10.1093/bioinformatics/btz096.


Motivation: The identification of differentially methylated regions (DMRs) among phenotypes is one of the main goals of epigenetic analysis. Although there are several methods developed to detect DMRs, most of them are focused on detecting relatively large differences in methylation levels and fail to detect moderate, but consistent, methylation changes that might be associated to complex disorders.

Results: We present mCSEA, an R package that implements a Gene Set Enrichment Analysis method to identify DMRs from Illumina450K and EPIC array data. It is especially useful for detecting subtle, but consistent, methylation differences in complex phenotypes. mCSEA also implements functions to integrate gene expression data and to detect genes with significant correlations among methylation and gene expression patterns. Using simulated datasets we show that mCSEA outperforms other tools in detecting DMRs. In addition, we applied mCSEA to a previously published dataset of sibling pairs discordant for intrauterine hyperglycemia exposure. We found several differentially methylated promoters in genes related to metabolic disorders like obesity and diabetes, demonstrating the potential of mCSEA to identify DMRs not detected by other methods.

Availability and implementation: mCSEA is freely available from the Bioconductor repository.

Supplementary information: Supplementary data are available at Bioinformatics online.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • CpG Islands
  • DNA Methylation
  • Promoter Regions, Genetic
  • Software*