Fig. 1
Meta-analysis workflow. All available scRNA-seq and snRNA-seq data were retrieved and downloaded from the Gene Expression Omnibus (GEO) or zenedo repository. Single cell RNA-seq and snRNA-seq samples were analysed separately, quality control metrics were measured and poor quality cells were filtered out in accordance. Then scRNA-seq samples and snRNA-seq samples were integrated independently. High resolution unsupervised clustering followed by visualisation of the expression of specific transcriptomic markers allowed to attribute each cluster a clear cell type (certain cell types were attributed to several clusters), or a cell type followed by « na » (i.e. not attributed) for the cells that did not show strong enough differenciation markers expression. Consensus signatures were computed using the FindAllMarkers() function in Seurat. Single cell RNA-seq and snRNA-seq were then integrated together to evaluate the matching between both datasets annotations. Finally, consensus signatures were used for cell type enrichments on previously published and annotated datasets.
