scRNA-seq analysis shows reduced UPR and increased oxidative stress in Atg7-deficient microglia. (A) UMAP of microglia sub-clustered from all CD45⁺ cells displaying eight distinct cell populations, including HM1–3, DAM, IFN-R (IFN-responsive microglia), TM, FTM, and proliferative microglia. (B) Dot plot showing the expression level of signature genes in each cluster. (C) Heatmap showing the scaled expression level of the cluster 6 (FTM) signature genes among all clusters. (D) Venn diagram showing the numbers and examples of overlapping feature genes between the FTM cluster and the MIMS-iron cluster in Absinta et al. (2021). (E) Frequencies of each cluster within the total microglia of each genotype. Each data point represents one mouse. Comparisons between 5xFAD and non-5xFAD samples were performed using two-way ANOVA; comparisons between Atg7^fl/fl and Atg7^ΔMG or Atg7^fl/fl-5xFAD and Atg7^ΔMG-5xFAD samples were performed using unpaired t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. (F–H) Volcano plots showing genes differentially expressed in HM from Atg7^fl/fl and Atg7^ΔMG mice (F), HM from Atg7^fl/fl-5xFAD and Atg7^ΔMG-5xFAD mice (G), and TM plus DAM from Atg7^fl/fl-5xFAD and Atg7^ΔMG-5xFAD mice (H). Genes with an average log₂ (fold change) >0.5 and an adjusted P value <0.05 are annotated in the figure. (I and J) Violin plots showing the module scores of the gene signatures in response to unfolded proteins gene set (I) and positive regulation of ROS metabolic process gene set (J). P value is calculated by stat_compare_means function in ggpubr package. **, P < 0.01; ****, P < 0.0001. Mean module score (black bar) is calculated by stat_summary function. UMAP: uniform manifold approximation projection.