Fig. 1: Overview of Vesalius’s cell state mapping strategy.

a Samples across conditions, technologies, and sections are likely to present a heterogeneous structure, and yet, to capitalize on the growing wealth of spatial omics data, it is crucial to accurately map cellular context across conditions. b To achieve cell mapping in heterogeneous spatial samples, Vesalius solves an LAP, which aims to minimize the overall cost. To account for spatial context during mapping, Vesalius leverages a multitude of biological features. c The total cost matrix can be constructed from a pair-wise summation of reciprocal similarity scores, including cell similarity, label similarity, niche similarity, composition similarity, territory similarity, and even custom similarity scores. d Vesalius maps cells across samples, across time, and across technologies. The total cost also provides a useful metric to define patient-level spatial similarities and allow for sample clustering. Source data are provided as a Source Data file.