Related Experiment Video
Updated: Dec 8, 2025
12:34
DNA Methylation: Bisulphite Modification and Analysis
Published on: October 21, 2011
106.0K
Detect differentially methylated regions using non-homogeneous hidden Markov model for bisulfite sequencing data
Yingyu Chen1, Chin Kiu Kwok2, Hangjin Jiang3
1Department of Statistics, Zhejiang University City College, Hangzhou, China.
Methods (San Diego, Calif.)
|September 19, 2020
Summary
We developed a novel Bayesian method, BSDMR, to detect differentially methylated regions in paired bisulfite sequencing data. This method improves accuracy and reduces false discoveries, aiding in disease research.
Area of Science:
- Genomics
- Epigenetics
- Bioinformatics
Background:
- DNA methylation is crucial in biological processes and diseases.
- Whole genome bisulfite sequencing (WGBS) enables detection of methylation patterns at single nucleotide resolution.
- Comparing normal and disease samples reveals aberrant methylation.
Purpose of the Study:
- To develop a novel Bayesian method for detecting differentially methylated regions (DMRs) from paired bisulfite sequencing data.
- To implement this method as an R package named BSDMR.
- To evaluate BSDMR's performance against existing methods.
Main Methods:
- Utilized a non-homogeneous hidden Markov model (HMM).
- Modeled spatial correlation between CpG sites.
- Incorporated relationships between methylation signals from paired normal and disease samples.
Main Results:
- BSDMR demonstrated robust performance, even with low sequencing read depth.
- Achieved lower false discovery rates compared to existing methods.
- Successfully identified DMRs in colon cancer data, validated by existing literature.
Conclusions:
- BSDMR offers a superior modeling strategy for analyzing paired WGBS data.
- The R package BSDMR is a valuable tool for identifying disease-associated methylation changes.
- This approach enhances our understanding of epigenetics in disease.
