DMRcaller: a versatile R/Bioconductor package for detection and visualization of differentially methylated regions in CpG and non-CpG contexts

Marco Catoni, Jonathan Mf Tsang, Alessandro P Greco, Nicolae Radu Zabet

Research output: Contribution to journalArticlepeer-review

17 Citations (Scopus)
257 Downloads (Pure)

Abstract

DNA methylation has been associated with transcriptional repression and detection of differential methylation is important in understanding the underlying causes of differential gene expression. Bisulfite-converted genomic DNA sequencing is the current gold standard in the field for building genome-wide maps at a base pair resolution of DNA methylation. Here we systematically investigate the underlying features of detecting differential DNA methylation in CpG and non-CpG contexts, considering both the case of mammalian systems and plants. In particular, we introduce DMRcaller, a highly efficient R/Bioconductor package, which implements several methods to detect differentially methylated regions (DMRs) between two samples. Most importantly, we show that different algorithms are required to compute DMRs and the most appropriate algorithm in each case depends on the sequence context and levels of methylation. Furthermore, we show that DMRcaller outperforms other available packages and we propose a new method to select the parameters for this tool and for other available tools. DMRcaller is a comprehensive tool for differential methylation analysis which displays high sensitivity and specificity for the detection of DMRs and performs entire genome wide analysis within a few hours.

Original languageEnglish
JournalNucleic Acids Research
Early online date9 Jul 2018
DOIs
Publication statusE-pub ahead of print - 9 Jul 2018

Fingerprint

Dive into the research topics of 'DMRcaller: a versatile R/Bioconductor package for detection and visualization of differentially methylated regions in CpG and non-CpG contexts'. Together they form a unique fingerprint.

Cite this