Fig. 2

Single-cell RNA sequencing of corpus callosum myeloid cells from naïve mice and mice on a five week cuprizone diet. a Schematic showing microdissections, dissociation, fluorescence activated cell sorting (FACS), single cell RNA sequencing, and bioinformatic workflow. b Unsupervised graph-based clustering and machine learning-based reclustering of dataset projected onto a Uniform Manifold Approximation and Projection (UMAP). Most microglia isolated from naïve mice (control cells, blue) segregate from microglia isolated from cuprizone fed mice (orange) demonstrating distinct microglial states. b, c Microglia segregated into three cuprizone associated microglia (CAM) clusters, three white matter associated microglia (WAM) clusters enriched with homeostatic cells, and two clusters with a mixture of cuprizone and control cells. d, e Kernel density estimates of genes related to lipid regulation (Apoe and Lpl) and inflammation (Clec7a, Axl, Ccl3, and Fth1) were projected onto UMAPs and were found to be enriched within CAM. f Hierarchical clustering of differentially expressed genes with Wards method produced four gene sets including 227 genes that could differentiate clusters. Legend represents average gene expression values with red and blue indicating higher and lower expression, respectively