Schematic of a typical CytoSPACE workflow. Given an ST dataset A and an annotated scRNA-seq dataset B, where the latter covers major cell types in A, CytoSPACE consists of the following key steps: (1) application of an existing ST deconvolution method (for example, Spatial Seurat or RCTD) to estimate cell type fractions in A using reference profiles from B; (2) estimation of the number of cells per spot in A; (3) sampling of B to match the inferred number of cells per cell type in A; and (4) alignment of single-cell and spatial transcriptomes (B → A) using shortest augmenting path-based optimization. The labels a 1, a n,..., d 1, d n denote individual single cells of cell type a, a,..., d,d, respectively.
