Fig. 4
Cuprizone-associated microglia are linked to inflammatory processes and reactive oxygen species production and resemble “disease associated microglia”. a Alluvial plot linking microglial clusters from naïve and five weeks post-cuprizone dataset to gene ontology terms. b Gene ontology terms projected onto UMAP of naïve and five weeks post-cuprizone microglia. ROS producer and ROS scavenger genes were compiled, summed, and projected onto a UMAP projection. c Unsupervised graph-based clustering onto a UMAP for cuprizone dataset combining clusters identifies five clusters: cuprizone associated microglia (CAM) 1 to 3, mixed, and white matter associated microglia (WAM). Safaiyan et al. 24 month old grey and white matter dataset contained four clusters: homeostatic (homeo), reactive white matter microglial (RWAM), and activated [55]. Frigerio et al. dataset with 3, 6, 12, and 21 month-old APPNL-G-F mice contained four clusters: disease associated microglia (DAM)1, DAM2, proliferative microglia (Prolif), and homeo [56]. d Cuprizone dataset, Safaiyan et al. dataset, and Frigerio et al. dataset were compared using the four gene sets derived from the differentially expressed genes from cuprizone dataset (this study). Data was normalized and red represents higher average expression while blue represents lower average expression