In dynamic tissues, spatial context is known to induce specific transcriptional changes, but the limited coverage and resolution of current spatial transcriptomic technologies ---and the absence of spatial context in higher resolution single-cell dynamics tools --- limit quantitative understanding of these cell-state transitions. This proposal will bridge this gap by integrating single-cell and spatial transcriptomic data to learn spatially informed gene programs and infer calibrated transition dynamics, returning reproducible statistics comparing replicates, conditions, and technologies. Through collaborations with domain experts, these open-source methods will enable rigorous discovery of niche-dependent processes across systems (e.g., limb development, intestine, fetal lung, tumor microenvironment) and will identify perturbation outcomes and spatial biomarkers of clinically relevant tissue states.

Who is involved: Akansha Gupta, Ruxandra Tonea, Joseph Sifakis, and Ruimin Zhang