Project Summary Single-cell immune profiling technologies, such as cytometry by time of flight (CyTOF) enable broad and comprehensive characterization of diverse immune cell-types. Moreover, such technologies are being increasingly applied in clinical settings to gain a holistic view of the immune system. Ex vivo stimulation is a common perturbation applied to immune cells and assayed through CyTOF, which elicits functional responses that may be clinically predictive. Such experiments generate single-cell measurements for a large number of cells, causing manual analysis to become time-consuming and biased towards studying immune cell-types and their functional responses that have already been well-characterized. Existing bioinformatics approaches for automating manual analysis are limited in that they 1) primarily focus only on partitioning cells into cohesive cell-populations, 2) need to be run independently per stimulation and 3) produce several immunological features encoding cell-type specific functional responses to stimulation that are not indicative of canonical immune signaling pathways. In this proposal, we introduce a fully automated approach for automating the analysis of multi-sample, multi-stimulation immune profiling data. In particular, we shall develop algorithms to efficiently identify clinically-predictive functional responses to stimulation in a scalable manner to enable analysis of large clinical cohorts under several stimulation conditions. Uncovered functional responses that are clinically predictive can be used to develop diagnostic tests or to design vaccines to elicit particular cellular responses.