Figure 2

Generation of an Integrated Lung Cell Atlas through integration of public data reveal fibroblasts and alveolar cells expressing ACE2 in the human lung. (a) Public data from single-cell RNA sequencing (scRNAseq) studies on the human lung was retrieved. Reyfman et al., Madissoon et al. and Human Cell Landscape control datasets were individually analyzed and batch-corrected with Seurat v3 anchor-based integration after filtering and normalization. Batch-corrected data representing each study was used to assemble an integrated dataset. This integrated dataset was annotated with the assistance of label transferred from Travaglini et al. data. (b) The Human Lung Integrated Cell Atlas represented in two-dimensional diffusion-based Manifold Approximation Embedding (dbMAP). dbMAP organizes the visualization to preserve as much of the original data structure as possible in a comprehensive way. In this representation, each cell is a point mapped to an embedding so that its x, y coordinates represent its relative transcriptional identity, i.e. its phenotypic signal. Cells are colored by their assigned cluster. (c) Visualization of ACE2 expression on the human lung dbMAP embedding. ACE2 is consistently expressed in alveolar cells and scattered in fibroblasts, being practically absent from other cell types. NK Natural Killer cell; DC Dendritic Cell; NA Not Assigned.