dmrseq

dmrseq detects differentially methylated regions (DMRs) from Whole Genome Bisulfite Sequencing data and provides rigorous statistical inference for genome-wide DNA methylation differences.

Key Features:

• Genome-wide detection: Scans the entire genome to identify regions with differential DNA methylation from Whole Genome Bisulfite Sequencing data.
• Statistical inference with pooled null: Employs a pooled null distribution to provide accurate statistical inference and uncertainty assessment, enabling analysis with as few as two samples per population.
• Modeling of methylation data: Fits a generalized least squares regression model to transformed methylation proportions and incorporates a nested autoregressive correlated error structure for region-level statistics.
• Performance and error control: Demonstrates improved specificity and sensitivity and accurate control of the false discovery rate (FDR) based on experimental data and Monte Carlo simulations.

Scientific Applications:

• Epigenetics studies: Detection and quantification of DNA methylation changes in epigenomic analyses using WGBS.
• Cancer research: Identification of DMRs associated with oncogenesis and tumor heterogeneity.
• Developmental biology: Characterization of methylation changes across developmental stages.
• Population and comparative studies: Comparison of DNA methylation profiles across populations or experimental conditions.

Methodology:

Fits generalized least squares regression to transformed methylation proportions with a nested autoregressive correlated error structure and uses a pooled null distribution for inference; performance evaluated using experimental data and Monte Carlo simulations.