Translating Geroscience to Clinic - Using ML for Insightful and Responsive Aging Biomarkers

Date of Award

Spring 1-1-2025

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computational Biology and Bioinformatics

First Advisor

Higgins-Chen, Albert

Abstract

With the global population aging at an unprecedented rate, there is a pressing need to shift from reactive, symptom-driven healthcare to gerocentric and preventive medicine. Addressing the complexities of aging and its associated diseases requires biomarkers that can quantify biological aging, predict age-related risks, and monitor responses to interventions. Biomarkers lie at the core of this transformation, serving as essential tools to understand, predict, and modulate the biological underpinnings of aging, ultimately enabling earlier and more precise interventions. This work introduces Systems Age, a novel framework leveraging DNA methylation (DNAm) data to measure aging heterogeneity across 11 physiological systems, including the heart, brain, immune system, and metabolism. By combining supervised and unsupervised machine learning, Systems Age provides system-specific scores that are predictive of diseases, mortality, and functional decline, while also identifying distinct aging subtypes that correlate with unique health outcomes. This systems-based approach enables clinicians and researchers to dissect the multidimensional nature of aging and design personalized strategies for managing age-related health. In addition, this thesis explores the responsiveness of DNAm biomarkers to interventions targeting aging biology. A meta-analysis of 51 longevity intervention studies compiled IN a database called TranslAGE-Responsive reveals that DNAm biomarkers respond differentially to pharmacological, dietary, and lifestyle interventions, reflecting their potential to dynamically monitor biological aging. These biomarkers provide critical insights into the effectiveness of interventions, demonstrating variability based on study population health, biomarker type, and intervention category. By addressing both the heterogeneity of aging and the ability to track its modulation, Systems Age and TranslAGE-Responsive highlight the transformative potential of geroscience to reshape healthcare. In this process we also lay the framework for general evaluation of these biomarkers using the TranslAGE framework using TranslAGE Prognostic and TranslAGE Reliability, showing the need to evaluate aging biomarkers from multiple perspectives. This research bridges the gap between aging biology and clinical application, laying the groundwork for a future where predictive and preventive medicine enhances healthspan and quality of life across an aging population.

This document is currently not available here.

Share

COinS