Inter-laboratory adaption of age estimation models by DNA methylation analysis-problems and solutions

Area of Science:

• Epigenetics
• Forensic Science
• Biotechnology

Background:

• DNA methylation-based age prediction models are emerging for biological age determination.
• Cross-laboratory validation is crucial for routine application of these epigenetic clocks.
• Previous models utilized specific CpG sites in genes like ASPA, EDARADD, PDE4-C, and ELOVL2.

Purpose of the Study:

• To assess the cross-laboratory accuracy of two published DNA methylation age prediction models.
• To evaluate the performance of these models on blood and buccal swab samples.
• To investigate methods for improving the accuracy and reliability of epigenetic age prediction.

Main Methods:

• Tested two established DNA methylation age prediction models using blood and buccal swab samples.
• Analyzed prediction accuracy by calculating the mean absolute difference (MAD) between chronological and predicted age.
• Retrained the prediction models and re-evaluated their accuracy.

Main Results:

• Initial testing showed higher MAD values (9.84 years for blood, 8.32 years for buccal swabs) than reported, suggesting inter-laboratory variability.
• Retraining the models significantly improved prediction accuracy, achieving MADs of 5.55 years (blood) and 4.65 years (buccal swabs).
• The study identified DNA methylation variances in certain CpGs as a likely cause for lower initial accuracy.

Conclusions:

• Published DNA methylation age prediction models may not perform as expected across different laboratories without standardization.
• Retraining and optimizing models can enhance prediction accuracy for biological age estimation.
• Effective DNA standards are essential for normalizing DNA methylation data, enabling reliable comparison and application of age prediction models.