MethylCal

MethylCal implements Bayesian calibration to correct PCR bias and produce calibrated methylation estimates from bisulfite amplicon sequencing for CpG sites within CpG islands (CGIs) and differentially methylated regions (DMRs).

Key Features:

• Bayesian calibration with INLA: Implements a Bayesian calibration framework using Integrated Nested Laplace Approximation (INLA) as introduced by Rue et al. (2009).
• Joint calibration: Analyzes all CpG sites within a CGI or DMR jointly rather than calibrating each CpG individually.
• PCR bias correction: Corrects preferential amplification effects (PCR bias) that affect methylation quantification in bisulfite amplicon sequencing.
• Calibration curves from controls: Derives calibration curves from standard controls with known methylation percentages to correct observed methylation levels.
• Outlier detection: Identifies outliers in methylation measurements to improve reliability of calibrated estimates.
• Differential methylation testing: Performs tests to detect hyper- and hypo-methylated samples for case–control comparisons.
• Alternative calibration methods: Includes implementations of calibration approaches from Warnecke et al. (1997) and Moskalev et al. (2011).
• Prediction of unobserved states: Provides calibrated predictions for unconsidered or unobserved methylation states based on the fitted model.
• Validation across assays: Validated across eight independent assays, including two CpG islands and six imprinting DMRs.

Scientific Applications:

• Methylation quantification: Calibration and correction of methylation measurements from bisulfite amplicon sequencing experiments.
• PCR bias mitigation: Reducing bias introduced by preferential amplification related to methylation state.
• Differential methylation analysis: Detection of hyper- and hypo-methylation in case versus control samples.
• Imprinting DMR analysis: Application to imprinting DMR assays, including validation across multiple imprinting regions.
• Clinical diagnostics: Applied in diagnostic contexts such as analyses for Beckwith-Wiedemann syndrome and celiac disease.
• Epigenetic research: Supporting precise measurement of methylation in epigenetic studies requiring accurate CpG-level estimates.

Methodology:

Uses a Bayesian calibration framework implemented with INLA (Rue et al., 2009) for joint analysis of all CpG sites within CGI/DMR, derives calibration curves from standard controls with known methylation percentages, includes outlier detection and tests for hyper-/hypo-methylation, and provides alternative calibration methods from Warnecke et al. (1997) and Moskalev et al. (2011).