scBrainScope: cross-species multidimensional brain atlas

Abstract

As one of the most evolutionarily complex and functionally diverse organs, the brain is characterized by its intricate structure, developmental alterations, and cellular heterogeneity. Although extensive brain atlases exist, there remains an urgent need for an integration platform that brings together gene expression profiles and analysis tools to explore features across species, brain regions, developmental stages, and pathological conditions. Thus, we developed scBrainScope, a detailed and interactive transcriptomic atlas of the brain. Our platform brings together single cell sequencing data from 135 species, covering 433 brain regions, 198 developmental stages, and 100 neurological diseases. In addition, we compiled 737 bulk RNA sequencing datasets from 275 species, along with 1154 spatial datasets of brain, spinal cord, and embryonic tissues. scBrainScope comprises six atlas modules (AtlasScope, RegionScope, TissueScope, SpaceScope, PathoScope, and AgeScope) and three analytical modules (sPandora, ePandora, and cPandora). Together, these tools enable researchers to investigate cell identity, gene programs, and spatial organization at multiple scales and dimensions. scBrainScope is freely available at http://8.142.154.29/scBrainScope or http://www.brainscopes.org, offering an interactive, data-rich resource for neuroscience, evolutionary biology, and translational medicine.

Graphical Abstract

Figure 1.

Comprehensive comparison between scBrainScope and existing brain-related transcriptomic databases. scBrainScope brings together 118.7 million single-cell transcriptomes (scRNA-seq) from 135 species, covering 433 brain regions, 100 brain disorders, and 198 developmental stages. It also includes 1154 ST sections from 16 species and 737 bulk RNA-seq datasets from 275 species. scBrainScope’s pan-dimensional analytical tools feature specialized modules for cross-species, cross-region, cross-development, and cross-disease analysis at the single-gene (sPandora), gene-set (ePandora), and cell-type (cPandora) levels. Compared to other databases like MAPbrain [5], Toti [43], SCAN [8], STOmicsDB [6], SpatialRef [44], ssREAD [45], SODB [46], Aquila [47], STellaris [48], SPASCER [49], STAB [50], SC2disease [51], SpatialDB [52], and scREAD [7] scBrainScope stands out with broader species coverage, greater data dimensionality, and fully integrated multimodal datasets and analysis tools. “NA” indicates missing data for certain metrics in the listed databases.

Figure 2.

Overview of scBrainScope single-cell transcriptomics modules (AtlasScope, RegionScope, PathoScope, and AgeScope) and user interface. (A) Diagrams showing the six transcriptomics atlas modules: AtlasScope, RegionScope, TissueScope, SpaceScope, PathoScope, and AgeScope. (B) Taking AtlasScope as example to showcase Single-dataset exploration workflow. Users start by selecting a species either by browsing classification or clicking on the species image [53, 54]. (C) Example of sPandora analysis workflow of four single cell transcriptomics modules. Users enter a single gene of interest to check its expression patterns across species and cell types, displayed as a heatmap. (D) Example of ePandora analysis workflow of four single cell transcriptomics modules. Users select a gene set [55] from 25 curated categories to explore cross-species expression patterns across cell types. The result is shown by a heatmap. (E) Example of cPandora analysis workflow of four single cell transcriptomics modules Users select a cell type to examine its proportion across species, visualization as stacked and conventional bar plots. (F) GeneLists function workflow of four single cell transcriptomics modules. Users select a dimension (cell types in AtlasScope module) to access DEG lists, with downstream analysis including enrichment (Motif [36], ToppGene [34, 35]), PPI networks [37, 56]. (G) Example of dataset selection in RegionScope, PathoScope, and AgeScope. Unlike AtlasScope, which begins with species-level or disease category-level dataset selection, these modules first require users to choose a broader category (e.g. species or disease category) before refining the selection to a specific brain region, developmental stage, or disease. The panel illustrates this two-step process using representative examples (RegionScope), while the available options extend beyond those shown.

Figure 3.

Overview of scBrainScope ST module (SpaceScope) and user interface. (A) Single-dataset exploration workflow in SpaceScope. Users start by picking a species and tissue type (such as brain, spinal cord, or embryo), then browse the available datasets and review the detailed metadata. The analysis panel offers multiple functions (e.g. spot visualization, spatial variable gene identification, enrichment analysis and spatial region annotations) [63, 64]. (B) sPandora workflow in SpaceScope. Users select a gene of interest and choose a category (Species, Devo, Tissue, Diz) to explore how that gene is expressed spatially across different conditions [, 65]. (C) ePandora workflow in SpaceScope. Users pick predefined gene sets [66] and examine their spatial expression across various categories [67]. (D) Users choose a cell type to explore its spatial distribution across different categories [20].

Figure 4.

Overview of scBrainScope bulk RNA-seq module (TissueScope) and user interface. (A) Single-dataset exploration workflow. Users select species by browsing the classification or clicking a species image to view the available bulk RNA-seq datasets, and detailed metadata [68]. (B) sPandora workflow on TissueScope. Users pick a single gene of interest and visualize its expression across 275 species with options to download the result in image, PDF, or CSV format. (C) ePandora workflow on TissueScope. Users can check a predefined gene set and examine its expression across species [55]. (D) GeneLists workflow on TissueScope. Users choose species classifications to access differently expressed gene lists and run enrichment analysis like pathway enrichment, motif enrichment [36], PPI networks [37, 56], and ToppGene enrichment [34, 35].

Figure 5.

Overview of scBrainScope analysis modules (sPandora, ePandora, cPandora) for pan-dimensional comparison of genes, cell types, and gene sets. (A) sPandora workflow: users enter a single-gene query and investigate its expression across all scBrainScope dimensions. The results are provided with heatmaps (AtlasScope, RegionScope, PathoScope, AgeScope), bar plots (TissueScope), or expression plots (SpaceScope) [63]. (B) ePandora workflow: users select the gene sets of interest and dimensions to view the result [55]. Users can explore the expression pattern of predefined gene sets across dimensions including species (AtlasScope, TissueScope), brain regions (RegionScope), spatial locations (SpaceScope) [76], disease conditions (PathoScope), and development stages (AgeScope). The results are visualized as heatmaps, bar plots or expression plots. (C) cPandora workflow: users select a cell type (e.g. neurons, microglia, astrocytes) or all cell types, and examine its proportion across species (AtlasScope), brain regions (RegionScope), spatial locations (SpaceScope) [67], disease conditions (PathoScope), or developmental stages (AgeScope). The corresponding results are displayed by stacked bar plots or expression plots.