(A and B) Microglia single cell trajectory inference describes the evolution of microglia activation from naïve to stroke condition. (A) Partition-based graph abstraction graph (PAGA, right) shows cluster connectivity of Louvain-clusters (left) with a threshold of 0.1. Nodes represent subsets, and thicker edges indicate stronger connectedness between subsets. The trajectories path 1 and 2 are selected based on the connectivity (edge width) in the PAGA after the root and end paths were defined based on marker gene expression. Clusters along the paths from root-to-end were merged since these are consecutive. Trajectory path 1 (dark gray): root (cluster 0), path 1 (clusters 6, 1, and 3 merged), and end (cluster 7); trajectory path 2 (black): root (cluster 0), path 2 (cluster 4, 2, and 5 merged), and end (cluster 7). (B) Barplot shows cell frequencies split by condition in trajectories path 1 and path 2. (C and D) Differential microglial gene expression analysis between root and end clusters along trajectory path 2. (C) Heatmaps show scaled expression along path 2 trajectory of microglial genes specifically regulated in Rag1−/− and compared in WT (left) and Rag1−/− (right) in path 2 (Wilcoxon test, adjusted p<0.05). For balanced representation of cells along the trajectory root and path 2, cells were randomly subset to 100 cells before plotting. (D) Top enriched pathways of microglial genes up- and down-regulated in Rag1−/− mice. (E–G) Microglia gene set expression in stroke condition. (E) Uniform manifold approximation and projection (UMAP) plots in path 2 and end cell clusters colored-coded by stroke condition (left), genotype (middle), and clusters (right). (F) Gene-gene correlation map of the correlated and anti-correlated genes extracted from clusters 2, 4, 5, and 7 identified 11 gene clusters. Due to low correlation, G8 was excluded of further analysis. (G) Dot plots show gene set expression distribution (G1–G11) of the stroke cell clusters (2, 4, 5, and 7).