Answers are grounded in this paper and cite their sources.
Spatial transcriptomics has made tissue architecture measurable at molecular resolution, yet most computational methods still identify spatially variable genes and infer biology only indirectly. We present SVP, a framework that directly maps spatially variable cell states, pathways and regulatory programs from spatial transcriptomic data. SVP combines location-level functional activity inference with global and local spatial autocorrelation and extends these analyses to bivariate spatial coupling across heterogeneous features. Across simulations and real datasets, SVP recovered cell-state-associated activity patterns at individual captured locations, detected spatially variable features with calibrated statistics, and scaled efficiently to large datasets. In pancreatic ductal adenocarcinoma, SVP identified a spatially co-organized hypoxic and extracellular-matrix-remodeling niche that co-localized with regions of low immune activity. It further captured functional maturation in the developing human heart and plaque-associated glial programs in Alzheimer's disease models. SVP enables direct analysis of functional spatial organization, moving beyond gene-centric descriptions of tissue structure.
No takes yet. Share an insight, caveat, or question.
Yu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: