MethylCal: Bayesian calibration of methylation levels

Area of Science:

• Epigenetics and Genomics
• Bioinformatics and Computational Biology

Background:

• Bisulfite amplicon sequencing is key for DNA methylation quantification but suffers from PCR bias.
• PCR bias leads to inaccurate methylation level estimations, necessitating calibration methods.

Purpose of the Study:

• To introduce MethylCal, a Bayesian calibration tool for accurate DNA methylation analysis.
• To address limitations of single CpG calibration by analyzing CpGs within CpG islands (CGIs) or Differentially Methylated Regions (DMRs) jointly.

Main Methods:

• Developed a Bayesian calibration approach (MethylCal) for joint analysis of CpGs within CGIs or DMRs.
• Tested MethylCal on eight independent assays, including CGIs and imprinting DMRs.
• Evaluated MethylCal's performance in clinical diagnostic cases for Beckwith-Wiedemann syndrome and celiac disease.

Main Results:

• MethylCal enables more precise modeling of methylation levels from standard controls.
• The tool provides accurate predictions for unobserved methylation levels, crucial for corrected degree derivation.
• Demonstrated benefits including outlier detection and improved identification of methylation changes in clinical samples.

Conclusions:

• MethylCal offers a robust solution for accurate DNA methylation quantification by correcting PCR bias.
• The method enhances the identification of methylation alterations in borderline clinical cases.
• MethylCal has the potential to influence clinical and treatment decisions in epigenetic-related diagnostics.