EpInflammAge: Epigenetic-Inflammatory Clock for Disease-Associated Biological Aging Based on Deep Learning

Area of Science:

• Aging research
• Computational biology
• Immunology

Background:

• Aging is characterized by epigenetic alterations and immunosenescence.
• Predicting biological age accurately is crucial for understanding health and disease.
• Existing epigenetic clocks have limitations in disease sensitivity.

Purpose of the Study:

• To develop EpInflammAge, an explainable deep learning tool for biological age prediction.
• To integrate epigenetic and inflammatory markers for enhanced accuracy and disease sensitivity.
• To bridge the hallmarks of aging: epigenetic changes and immunosenescence.

Main Methods:

• Utilized deep neural networks for tabular data analysis.
• Integrated epigenetic (DNA methylation) and inflammatory (cytokine) data from participants.
• Employed open-source epigenetic data to generate synthetic inflammatory biomarkers.
• Trained an age estimation model using combined datasets.

Main Results:

• EpInflammAge achieved competitive performance against 34 epigenetic clock models.
• Demonstrated a mean absolute error of 7 years and a Pearson correlation of 0.85 in healthy controls.
• Showcased robust sensitivity across multiple disease categories.
• Explainable AI identified feature contributions to age prediction.

Conclusions:

• Combining epigenetic and inflammatory profiles enhances biological age prediction sensitivity to disease.
• EpInflammAge offers a promising tool for both research and clinical applications.
• The explainable nature of the tool aids in understanding predictive factors.