Background & Rationale.
Chronic diseases and neurodegenerative disorders reflect complex interplay among inherited risk, baseline physiology, epigenetic aging, and everyday behavior (e.g., sleep-wake timing, activity rhythms). Conventional analyses often overlook high-dimensional, cross-modal dependencies. We propose an integrated, multimodal framework-grounded in quantitative modeling, machine learning, and biological validation-to uncover high-order associations among genetics, epigenetics, physiology, and behavior, enabling earlier risk stratification and individualized interventions.
Research Questions.
1. To what extent does combining genetic, epigenetic, physiological, and behavioral features improve prediction of dementia and chronic disease onset versus single-modality models?
2. Which cross-modal interactions (e.g., gene×circadian rhythm, epigenetic age×lifestyle) provide the strongest prognostic signals for metabolic, cardiovascular, psychiatric, and neurodegenerative disorders?
3. Can we trace trajectories from resilience to pathological aging in a unified multimodal feature space?
4. What mechanisms underlie dementia prevention in individuals with rejuvenating lifestyle habits compared to those without?
Objectives.
* High-Performance Prediction: Develop multimodal risk models (Cox survival, deep networks) for early disease detection and dementia clock construction.
* Epigenetic Integration: Identify genetic, environmental, and biological determinants of accelerated epigenetic aging and link them to disease outcomes.
* Personalized Intervention: Design individualized strategies triggered by multimodal risk profiles, validated in occupational health contexts (e.g., shift work).
* Cancer/TME Insights: Integrate genetic and biomarker data to uncover regulatory mechanisms of tumor metabolism and immune microenvironment.
* Brain Resilience: Identify neuroimaging and exposomic markers of preserved cognition in resilient aging.
