Fig. 3
Subclustering of snRNAseq data reveals region- and genotype-specific subpopulations of microglia. A Subcluster UMAP of microglia nuclei transcriptomes showing two main clusters denoted as A and B. B Number of microglia per condition in each cluster. C Microglia UMAPs indicating cells from each condition (dark blue). D Boxplots of distributions of log2 gene expression for each module of co-expressed genes across conditions. Standard boxplot indicates box as median with upper and lower quartiles and whiskers as range excluding outliers (beyond 1.5*interquartile range from the upper/lower quartiles). E Dot plot indicating number of DEGs in each gene module (dot size) and average LFC (colour scale) for three DE tests: WT: EC vs. DCN, EC: APP vs. WT, DCN: APP vs. WT. F UMAPs showing aggregated expression of genes in the top 3 largest gene modules. Colour scale indicates expression in each cell relative to the cell with the highest expression of that module. G UMAPs of expression of selected top marker genes from each module showing percentage of maximum expression. H, I Gene Ontology analysis of biological processes enriched in genes in Mg.G2 (App KI-enriched; H) and Mg.G3 (DCN-enriched; I) compared to all detected microglial genes. Bar plot colour scale indicates -log10 p-value. J FISH assay to determine co-expression of IFN-regulated gene and microglia marker, Oasl2 (green) and Hexb (red) respectively, in the EC (top panel) and DCN (bottom panel) of WT and App KI tissues. White closed circles indicate Hexb-positive microglia expressing Oasl2 while broken circles indicate Hexb-positive cells that lack Oasl2 expression. Scale bar at 10 µm. K As (J), but for the colocalisation of Ifi204 and Hexb in the EC and DCN of WT and App KI tissue. Closed circles indicate Hexb-positive microglia that express Ifi204 while broken circles indicate microglia that do not express Ifi204. Scale bar at 10 µm. (L-M) Percentage of Hexb-positive cells that express Oasl2 (N = 3 subjects, n = 3 sections/subject; L) and Ifi204 (N = 3 subjects, n = 3–4 sections/subject; M). Graphs indicate mean ± SEM. Two-way ANOVAs were performed on within-subject means to test for main effects of region and genotype and interaction effects, followed by Tukey HSD post hoc tests (¤). ** p < 0.01