Novel Machine Learning Methods for Immunosenescence and Aging Research

This proposal aims to develop advanced statistical and computational methods for analyzing immune data to enhance our understanding of the immune system’s role in aging and age-related diseases. Aging is accompanied by significant changes in the immune system, increasing susceptibility to infections, chronic inflammation, and other health challenges. Understanding the heterogeneity of immune profiles and their associations with aging outcomes holds great promise for predicting disease risk and identifying therapeutic targets. Large-scale studies such as the Health and Retirement Study (HRS) provide invaluable data on the elderly population. However, immune data obtained from flow cytometry present unique analytical challenges. These data are compositional, highly skewed, and prone to substantial measurement errors, rendering standard analyses unreliable. Existing methods for supervised and unsupervised analysis of immune data frequently fall short in adjusting for covariates, identifying key immune features, capturing nonlinear relationships, and integrating multiple data sources, resulting in significant gaps in our understanding of immune aging. This proposal addresses these challenges through innovative methodologies. In Aim 1, we will develop a robust nonparametric framework to denoise immune cell frequency data. This framework is free from distributional assumptions and adaptable to diverse data types, enhancing the accuracy of subsequent analyses. In Aim 2, we will create a model-based clustering framework to identify immune subgroups, with a special focus on adjusting for covariates and identifying key drivers of heterogeneity between clusters. In Aim 3, we will develop novel semi-parametric methods to integrate multiple sources of immune biomarkers and associate them with aging outcomes, emphasizing biological interpretability and feature selection. In Aim 4, we will build an open- source software package to ensure the accessibility and wide dissemination of these methods. Motivated by and applied to HRS data, these methods aim to uncover immune signatures in the elderly and clarify their relationship with age-related outcomes. The research will deliver powerful tools for immune data analysis and transformative insights into the interplay between immunity and aging. Project Number: 1R01AG097771-01 | Fiscal Year: 2026 | NIH Institute/Center: National Institute on Aging (NIA) | Principal Investigator: Gen Li | Institution: UNIVERSITY OF MICHIGAN AT ANN ARBOR, ANN ARBOR, MI | Award Amount: $298,419 | Activity Code: R01 | Study Section: Analytics and Statistics for Population Research Panel A Study Section[ASPA] View on NIH RePORTER: https://reporter.nih.gov/project-details/11274107