Extended Data Fig. 1: Single-cell mouse hippocampus integrated database: clusters and cell types prediction.

a, Single-cell mouse hippocampus datasets used to create the integrated database. Notice that studies generating individual datasets present relevant biological (age of mice, hippocampal region) and/or methodological (dissociation methods, chemistry, platform, enrichment strategies) differences among them. b, UMAP representation of 8 integrated hippocampus scRNA-seq datasets labelled by dataset. c, Cluster analysis revealed 15 transcriptionally distinct clusters (clustering resolution = 0.1). d, Integrated hippocampus database coloured by cell type prediction. e, Deep neural network multiclass model trained on a comprehensive database²⁹ at “subclass level”. Top, Accuracy and loss value after each epoch for training and validation data; Bottom, confusion matrix showing cell prediction for validation data. f, We performed cross-validation by removing each individual dataset one at a time and running the prediction and clustering using the others, then calculated the overall prediction efficiency. This showed the accuracy, sensitivity, and specificity of our model. Data are shown as box plot (25–75 percentile with median) with min to max whiskers excluding outliers. g, Expression levels for canonical astrocytic markers in the integrated hippocampus database.