DISMISS: detection of stranded methylation in MeDIP-Seq data

Area of Science:

• Genomics
• Epigenetics
• Bioinformatics

Background:

• DNA methylation regulates gene expression and chromatin structure.
• Methylated DNA immunoprecipitation sequencing (MeDIP-Seq) identifies DNA methylation but struggles with strand-specific signals.
• Asymmetric DNA methylation is understudied compared to symmetric CG methylation.

Purpose of the Study:

• Introduce DISMISS, a novel software package for detecting strand-associated DNA methylation from MeDIP-Seq data.
• Enable the analysis of asymmetric DNA methylation, which is prevalent in some invertebrate genomes.
• Improve the resolution of DNA methylation patterns in MeDIP-Seq datasets.

Main Methods:

• Developed DISMISS software to process MeDIP-Seq reads and identify strand-specific methylation.
• Utilized MeDIP-Seq datasets from Apis mellifera (honeybee) for validation.
• Compared DISMISS performance against bisulfite sequencing (BS-Seq) and MACS2.

Main Results:

• DISMISS accurately identifies strand-specific DNA methylation in MeDIP-Seq data, comparable to BS-Seq.
• Achieved 80% accuracy in predicting methylated cytosines in honeybee samples compared to BS-Seq.
• DISMISS correctly detects the origin of DNA methylation at splice junctions and outperforms MACS2 for various genomic loci.

Conclusions:

• DISMISS software overcomes limitations in detecting asymmetric DNA methylation from MeDIP-Seq data.
• Facilitates sensitive examination of MeDIP-Seq datasets, particularly for genomes with low or high asymmetric methylation.
• DISMISS is a valuable tool for studying diverse eukaryotic genomes and their methylation patterns.