Modeling Immune Data for Heterogeneity and Patient Outcomes

Lunch will be served at 11:45 AM

Abstract:  The immune system is central to infection, transplantation, cancer therapy, and many other areas of medicine, yet the data used to characterize immune responses are complex, high-dimensional, and heterogeneous. In this talk, I will present computational frameworks that address two key challenges: capturing heterogeneity in immune data and predicting patient outcomes in clinical studies where sample sizes are inherently small. I will highlight approaches that combine statistical models with optimal transport to characterize immune heterogeneity, and with deep learning to enable patient outcome prediction. Together, these methods show how computation can extract insight from immune measurements and connect them to clinically meaningful endpoints. I will conclude by briefly discussing opportunities to extend these approaches to multi-modal data, with the aim of providing a more complete picture of the immune system. 

Bio: Dr. Lynn Lin is an Associate Professor in the Department of Biostatistics and Bioinformatics at Duke University, and Associate Director of the Quantitative Sciences Core in the Duke Center for AIDS Research. She is also affiliated with the Duke Center for Human Systems Immunology. Trained in statistics and computational biology, Dr. Lin’s research develops statistical and machine learning methods for analyzing complex, high-dimensional biomedical data. Her work spans single-cell and multi-omics analysis, longitudinal modeling, and integrative frameworks that connect molecular and clinical data. Beyond methodological advances, she collaborates widely across disciplines, with applications in immunology, infectious diseases, oncology, and precision medicine.