Methodological challenges in constructing DNA methylation risk scores.

Area of Science:

• Epigenetics and Genomics
• Complex Trait Analysis
• Biomarker Discovery

Background:

• Polygenic approaches using genetic risk scores (GRS) explain more complex trait variance than single-variant methods.
• GRS are established tools for risk prediction, association, and interaction studies with genotype data.
• Growing interest in applying GRS principles to DNA methylation data, termed methylation risk scores (MRS).

Purpose of the Study:

• To review the adaptation of existing GRS methodologies to DNA methylation data.
• To highlight the potential applications of MRS in various biological and clinical research areas.
• To identify and discuss the challenges encountered when developing and applying MRS.

Main Methods:

• Direct transfer of established GRS methodologies to DNA methylation data.
• Exploration of MRS applications including biomarker analysis, association studies, and risk prediction.
• Identification of confounding factors (e.g., age, tissue type) impacting MRS development.

Main Results:

• Most GRS approaches are transferable to methylation data, enabling the creation of MRS.
• MRS offer potential for environmental exposure biomarker identification and association analyses.
• Challenges arise in obtaining appropriate external weights for MRS due to methylation data's sensitivity to confounders.

Conclusions:

• Methylation risk scores (MRS) represent a promising extension of polygenic approaches to epigenomic data.
• MRS can serve as valuable tools for disease risk prediction, biomarker discovery, and complex trait analysis.
• Addressing confounding factors is crucial for the robust development and application of MRS.