Fig. 4: Pathway imputation of a mouse brain dataset.
From: Accurate imputation of pathway-specific gene expression in spatial transcriptomics with PASTA
A Overview of the mouse brain cells and the predicted pathway expression. We grouped the cells into three main clusters. B Summary of correlation between the predicted and the true pathway expressions over 24 pathways. Each Dot represents a correlation value. The center bar represents the median value, and the upper/lower bound of the box shows the 25%/75% quartile with whiskers extending to the minimum and maximum values. PASTA shows significant better predictions compared to other methods except for Tangram (Two-sided Wilcoxon rank sum test, P = 0.0002 compared to StAI, P < 0.0001 for the others. *** indicates significant comparisons). C UMAP of the layer cells. D Pseudo-development time of the layer cells (111 L2 cells, 2909 L2/3 cells, 1789 L4/5 cells, and 438 L6 cells) estimated using pathway expression matrix from PASTA and Tangram. Each Dot represents a pseudo-development time value, and the center bar represents the median value. E K-means cluster using the pathway expression matrix from PASTA. F Major cell distributions of the K-means clusters using the pathway expression matrix from PASTA, Tangram, Seurat. Astro-Epen astrocyte and ependymal-like cells, GABA Gamma-aminobutyric acid, Glut glucose transporters, CB cerebellum, IT-ET inhibitory-excitatory, IMN immature neurons, DG dentate gyrus, OEC olfactory ensheathing cell, OPC oligodendrocyte progenitor cells. Source data are provided as a Source Data file.
