We downloaded the raw single-cell RNA-seq count matrices (deposited at GEO, accession code GSE98969) comprising immune cells isolated from the mouse brain in wild-type (WT) and AD mice19. The complete dataset with all samples consists of 37,248 cells. We filtered out ERCC spike-ins before computing the quality metrics of all cells. We then excluded 12,053 cells with less than 500 UMI counts and 11,065 genes, which were not expressed. We subsequently normalized by library size with target sum 10,000 counts (CPM normalization) and log+1 scaled. Following the analysis of Keren-Shaul et al.19, we selected the samples of six-month-old mice from AD and WT, which have not been sorted by brain region, resulting in 9,196 cells. It must be noted that Keren-Shaul et al. reported 8,016 cells when they first annotated immune cells in 6-month-old mice (see Fig. 1 in Keren-Shaul et al.). We evaluated batch effects based on the clustering results and visual inspection of the UMAP plots, where none of the samples clustered separately in any of the clusters, which is, in this case, sufficient to obtain cell types.
