Application of CAME to human and mouse scRNA-seq data of brain cells. (A) The predicted cell-type probabilities for each cell (each column) in the mouse brain scRNA-seq data. A maximum of 50 cells was subsampled from each type for visualization. The gene expressions of the human brain were taken as the reference. Each row indicates a cell type in human data. (OPC) Oligodendrocyte precursor cells, (SMC) smooth muscle cell, (VLMC) vascular and leptomeningeal cell. (B) The top homologous DEG expressions of oligodendrocytes and (predicted) OPCs in human and mouse data, including several marker genes reported by previous literature (collected from CellMarker; colored by red or blue). (C) Cross-species alignment of the gene embedding output by CAME, where each dot represents a gene, and each edge indicates the homology between a pair of genes. Genes shared between species are colored by light blue (human) or pink (mouse), and the other genes are colored by dark blue (human) or dark red (mouse). (D) The UMAP plots of gene embeddings showing the average expression patterns (z-scored across cell types for each gene) of four cell types (excitatory neurons, inhibitory neurons, oligodendrocytes, OPCs) of human and mouse brains, where the color of each dot is scaled by the expression level of that cell type in the gene. (E) Abstracted graph of the heterogenous cell–gene graph; each node represents a cell type (pink) or a gene module (light blue). The size of a node is scaled by the number of single cells in that type or the number of genes in that gene module. The width of an edge is scaled by either the normalized mean expression levels of a cell type in the connected gene module or the conservancy of inter-species gene modules based on the gene embeddings learned by CAME. (F) Gene modules detected by joint module extraction of genes from humans (above) and mice (below).
