Methods for discovering genomic loci exhibiting complex patterns of differential methylation

Area of Science:

• Genomics
• Epigenetics
• Bioinformatics

Background:

• Cytosine methylation is crucial for gene regulation in eukaryotes.
• Sodium bisulfite conversion enables genome-wide methylation quantification via high-throughput sequencing.
• Identifying methylation loci is key for understanding genomic regulation.

Purpose of the Study:

• To extend the segmentSeq R package for identifying methylation loci from multi-condition high-throughput sequencing data.
• To develop a statistical model for robust methylation analysis using an empirical Bayesian framework.
• To enable differential methylation analysis between multiple experimental conditions.

Main Methods:

• Utilized the segmentSeq R package and developed the baySeq R package.
• Implemented a statistical model accounting for biological replication and non-conversion rates.
• Applied an empirical Bayesian framework to compute posterior likelihoods of methylation.

Main Results:

• Successfully identified methylation loci in complex datasets, including Arabidopsis Dicer-like mutants.
• Revealed novel behaviors and antagonistic relationships between Dicer-like proteins at methylation loci.
• Demonstrated superior power in detecting differential methylation compared to existing methods in simulation studies.

Conclusions:

• The developed methods allow comprehensive analysis of methylome high-throughput sequencing data across diverse experimental conditions.
• Enables accurate identification of methylation loci and likelihood evaluation for downstream genomic characterization.
• Facilitates the characterization of diverse differential methylation patterns.