Fig. 4

Integration of different spatial technologies to undercover biological patterns using Matisse. A Spatial map of the osmFISH dataset, where every cell is colored by an RGB representation of 3 first dimensions of a UMAP representation. B Spatial map of the osmFISH dataset, where every cell has been assigned to a cluster found by hierarchical clustering. C Mean expression of every gene in the 20 clusters found de novo in the osmFISH dataset. D Correspondance between the predicted de novo clusters and cell types assigned to each cell in Codeluppi et al. [6]. E Gradient defined in the osmFISH dataset, where the white line represents the origin of the gradient and cells are colored depending on the minimum distance from itself to the origin line. Blue colors represent smaller distances, whereas yellow colors represent bigger distances. F Kernel density estimation of the frequency of 4 specific clusters (4,8,11 and 13) along the one-dimensional axis defined in E