Fig. 1

scRNA-Seq enables high-resolution interrogation of pancreatic cell-type specific expression. A Schematic of human pancreas and pancreatic cell types, contrasting bulk islet sequencing and single cell RNA sequencing. B Browser view RNA-Seq coverage of a representive acinar cell marker (REG1A). Tracks are depth normalized coverage tracks in units of counts per million (CPM). 3'-biased protocols (Drop-Seq/CEL-Seq) and full length (SMART-Seq) protocols are contrasted, as well as libraries from bulk, sorted cells and single cells. C Sequencing depth per cell in CEL-Seq (grey) and SMART-Seq (red) protocols. D Cumulative read count per gene in 2,000 CEL-Seq (grey) and SMARTseq2 (red) libraries. Numbers inset are the number of genes detected at a cumulative threshold of 1 read or 1,000 reads. E Number of cells in each study analyzed broken down by cell type (left) and the total number of cells by type. F Uniform Manifold Approximation and Projection (UMAP) of transcript abundances. At left: cells color coded by the study of origin, showing that data do not cluster by study. At right: cells color coded by cell type, showing that data cluster by cell type