The immune system's function and health involve complex interactions, proportions, and activation states of cells. Understanding human immune responses requires standardized approaches from subject recruitment to data analysis. Immune Health provides an infrastructure for recruiting subjects, processing biospecimens, running assays, and integrating data, with a standardized process for collecting and preparing blood specimens for single-cell analysis by cytometry.

High-throughput single-cell cytometry data are vital for understanding the immune system's role in diseases and treatment responses. Traditional annotation methods face challenges in scalability and accuracy. To address this, we propose the cytometry masked autoencoder (cyMAE), an automated solution for immunophenotyping, including cell type annotation. The cyMAE model uses Masked Cytometry Modeling (MCM) to learn relationships between protein markers without prior information and is fine-tuned for specialized tasks. Validated across multiple cohorts, cyMAE accurately identifies antibody co-occurrence patterns, provides interpretable cellular immunophenotyping, and improves prediction of subject metadata. It enhances predictions for cell type annotation, SARS-CoV-2 infection, secondary immune response, and COVID-19 infection stages. CyMAE significantly advances immunology research, enabling better prediction and interpretation of cellular and subject-level phenotypes in health and disease.

Dr. Dokyoon Kim is an Associate Professor of Informatics and the Director of the Center for AI-Driven Translational Informatics (CATI). He is also the Associate Director of Informatics at Immune Health. His research focuses on integrating multi-modal data, including 'omics data, environmental data, imaging, and Electronic Health Records (EHR) phenotype data. Dr. Kim's work spans theoretical and applied projects, developing methods to combine multi-omics data, predict clinical outcomes, and integrate genomics with imaging data. His long-term goal is to advance precision medicine through sophisticated data integration methods using AI. Dr. Kim started his academic career in 2016 at Geisinger Health System and later joined the University of Pennsylvania, where he continues to innovate in informatics research, enhancing patient outcomes by bridging data science and clinical practice.