BCurve

BCurve detects differentially methylated CG sites and regions using a Bayesian smoothing-curve method to model local methylation correlation and adjust for covariates in BS-seq and methylation microarray data.

Key Features:

• Unified Analysis Framework: Supports analysis of both bisulfite sequencing (BS-seq) and methylation microarray data while accommodating platform-specific characteristics.
• Bayesian Curve Credible Bands: Implements Bayesian smoothing-curve credible bands to model correlation between nearby CG sites and quantify uncertainty.
• Covariate Adjustment: Incorporates covariates such as sex and age into the methylation models to control for their effects.
• Between-Sample Variability Modeling: Accounts for variability between samples within the Bayesian framework to improve differential methylation inference.
• Simulation Capabilities: Provides simulation tools for generating BS-seq and microarray methylation datasets for method comparison and benchmarking.

Scientific Applications:

• Differential methylation mapping: Identification of differentially methylated CG sites and regions across experimental conditions in BS-seq and microarray studies.
• Epigenetic and gene regulation studies: Linking methylation patterns to gene regulation and epigenetic mechanisms.
• Developmental biology: Investigating methylation changes associated with developmental processes.
• Disease research and biomarker discovery: Characterizing methylation alterations in disease pathogenesis and identifying potential epigenetic biomarkers.
• Method benchmarking: Using simulated BS-seq and microarray data to compare and validate differential methylation methods.

Methodology:

Uses a Bayesian smoothing-curve approach with credible bands to model correlation between adjacent CG sites, incorporates covariates (e.g., sex and age), accounts for between-sample variability, and includes simulation of BS-seq and microarray data.