We present PathwayAge, a biologically informed, machine learning–based epigenetic clock that integrates pathway-level biological information to predict biological age and quantify disease-related aging acceleration. The framework is designed to operate across multiple tissues (e.g., blood and skin) and across different omics layers, including DNA methylation and transcriptomics[1]. Unlike traditional CpG-level clocks, PathwayAge models aging trajectories at the functional pathway level (GO/KEGG), thereby enhancing biological interpretability and enabling the identification of disease-specific aging patterns across tissues and molecular modalities.
The core algorithm is currently available on a separate GitHub repository. If you are conducting research and would like access, please feel free to reach out via email pan.Li.1122@proton.me.
This repository contains downstream analyses and validation scripts organized into four major components:
-
1-Evaluation /: Model performance evaluation: MAE, R², Spearman correlation, Internal and external validation. -
2-SkinFeatureRanking /: Feature importance ranking: Pathway contribution analysis, Stability assessment, Identification of aging-associated pathways. -
3-DiseaseControlStatisticalAnalysis /: Statistical comparison between disease and control groups: Logistic regression, Odds ratios (OR), Confidence intervals (CI), Age acceleration analysis. -
4-SurvivalAnalysis /: Clinical outcome evaluation: Survival analysis, Cox regression.References
[1] Li, P., et al. Decoding disease–specific ageing mechanisms through pathway-level epigenetic clock: insights from multi-cohort validation. eBioMedicine. Volume 118, 105829.