Age Prediction Using DNA Methylation Heterogeneity Metrics.

Area of Science:

• Epigenetics and Genomics
• Computational Biology
• Aging Research

Background:

• Genomic DNA methylation patterns dynamically change during development and aging.
• Epigenetic clocks (eAge) use DNA methylation to estimate biological age and disease risk.
• High-throughput sequencing reveals methylation heterogeneity linked to biological complexity.

Purpose of the Study:

• To systematically evaluate metrics for assessing within-sample heterogeneity (WSH) in DNA methylation.
• To construct and compare human blood epigenetic clock models using reduced-representation bisulfite sequencing (RRBS) data.
• To identify robust metrics for developing accurate and efficient epigenetic clock models.

Main Methods:

• Utilized RRBS data from human blood samples.
• Evaluated five distinct scores for assessing DNA methylation variability and WSH.
• Developed and compared epigenetic clock models based on WSH metrics and average methylation levels.

Main Results:

• A model based on a DNA methylation erosion metric achieved the best performance, with a Mean Absolute Error (MAE) of 3.686 years.
• A region-based model using average methylation levels was more efficient but less accurate than the WSH-based model.
• The WSH-based model requires analysis of only a few short genomic regions.

Conclusions:

• Metrics assessing DNA methylation erosion are effective for building accurate epigenetic clocks.
• A WSH-based epigenetic clock model using targeted sequencing offers a potentially useful tool for aging research.
• This approach could enable the development of reduced epigenetic clocks for practical applications.