Fig. 1

ONTraC captures spatially continuous changes in the tissue microenvironment. a A schematic of the ONTraC workflow. b–e Analysis of simulated_dataset_1. b UMAP (left) and spatial (right) distribution of cell types in simulated dataset 1. c Spatial distribution of ground truth (left) and ONTraC output (right) cell-level NT scores. d Cell type density along the reconstructed niche trajectory. e Bar plots showing the Spearman correlation between the ground truth and the ordering scores predicted by different methods. ONTraC, SpatialPCA, and SpaceFlow were applied only to the whole dataset, whereas the other methods were applied to the whole dataset, lineage 1, and lineage 2, respectively. f–h Analysis of the mouse motor cortex MERFISH data. f Spatial distribution of cell types on representative slices. g Spatial distribution of cell-level NT scores on representative slices. h Cell type density along the reconstructed niche trajectory