Adapting Machine Learning Approaches to Identify a Robust Measure and Key Risk and Protective Factors of Change in Allostatic Load across the Adult Lifespan

About This Grant

Chronic diseases—such as cardiovascular disease, diabetes, and Alzheimer’s Disease and Related Dementias —are leading causes of morbidity and mortality in U.S. adult, place a significant burden on the healthcare system and economy. Given that most chronic diseases are preventable, it is crucial to identify a reliable measure that can effectively detect early signs of chronic disease. Allostatic load (AL), an indicator of cumulative physiological “wear and tear,” has shown promise as a pre-clinical marker. However, consensus on its measurement in population datasets is lacking, and there is a lack of longitudinal evidence capturing changes in AL over time. This project aims to address three key objectives: 1) Identify the optimal AL measure and develop a DNA methylation (DNAm) surrogate of AL that best predicts health outcomes in two population datasets; 2) Examine AL change patterns across different life stages; and 3) Identify key risk and protective factors influencing changes in AL across life stages using advanced supervised machine learning techniques. Findings will establish essential guidelines for AL scoring in two nationally representative datasets: (1) the Health and Retirement Study (HRS) and (2) the National Longitudinal Study of Adolescent to Adult Health (Add Health). Findings will also provide optimization protocols for AL measurement in other biomarker-rich population datasets. Additionally, the project will highlight key risk and protective factors and critical intervention periods, informing the design of tailored intervention programs and policies to support healthy aging. The proposed project integrates an interdisciplinary research program, mentorship, education, and apprenticeships to achieve the following core training objectives: 1) Gain expertise in life-span development with specific focuses on mid and later life; 2) Develop expertise in human population biology, biospecimen collection techniques, and the epigenetics of health and aging; 3) Gain expertise in working with field-based longitudinal nationally representative datasets rich in biosocial data; 4) Develop skills in measurement theory and data science; 5) Engage in professional development. Training in these domains will support the candidate’s long-term career goal of becoming an independent investigator who integrates social science and human biology with the emerging field of data science to examine how life experiences become biologically embedded, influencing disease etiology and contributing to the development of chronic diseases. The candidate's existing foundation and research experience position them well to pursue this critical line of inquiry, and the K01 award is crucial for providing the career development needed to pursue this critical inquiry. Receiving the K01 award would significantly contribute to their intellectual development, research skills, and readiness to become an independent scholar.