methscore: a comprehensive R function for DNA methylation-based health predictors.

Area of Science:

• Epigenetics
• Genomic Medicine
• Computational Biology

Background:

• DNA methylation predictors are valuable for epidemiology and personalized medicine.
• Generating predictors from original sources is complex and time-consuming.
• Array and batch effects hinder predictor comparability across studies.

Purpose of the Study:

• To integrate 158 published DNA methylation predictors into a single R function.
• To provide a method (ref-RCP) to mitigate batch and array differences.
• To improve the precision and comparability of epigenetic predictor estimations.

Main Methods:

• Integrated 158 published DNA methylation predictors into an R function.
• Utilized classical and principal component-based methods for predictor generation.
• Developed a modified RCP method (ref-RCP) for normalizing DNA methylation data.

Main Results:

• The R function produces 158 predictors for chronological age, biological age, exposures, lifestyle, and serum proteins.
• The ref-RCP method normalizes input DNA methylation data to reference data.
• Evaluations demonstrated improved estimate precision and cross-study comparability.

Conclusions:

• The integrated R function simplifies the use of DNA methylation predictors.
• The ref-RCP method enhances the reliability of epigenetic estimations.
• The ENmix package offers a valuable tool for epigenetics research and personalized medicine.