Uncovering cross-signal interactions in single-cell and spatial omics

CRISP characterizes cross-signal interactions in single-cell gene expression by matching probabilistic signal-specific projections, recovering diverse interactions across simulations and four transcriptomic datasets.

• Why it matters: Existing approaches can disentangle signals, but they struggle to characterize how multicellular signals interact, where those interactions occur in tissue, and which genes may drive them.
• What they did: CRISP uses correspondence between multiple probabilistic signal-specific projections to quantify interactions, localize interacting subpopulations and associated gene groups, and refine labels across simulations and four single-cell/spatial transcriptomic datasets.
• The result: It distinguishes healthy from fibrotic liver and localizes loss of zonation-position coupling to individual fibrotic hepatocytes, resolves asymmetric TGF-{beta} response–EMT–cell-cycle coupling, flags transcriptionally delayed hepatocytes in Plasmodium liver infection, and can refine labels and reveal condition-associated genes.

AI summary of the abstract · may contain errors · not medical advice · read the paper.