Fig. 4
From: EmptyDropsMultiome discriminates real cells from background in single-cell multiomics assays

Biological comparison of EmptyDropsMultiome with CellRanger-arc on sample A (top) and sample B (bottom). A Uniform Manifold Approximation and Projection (UMAP) of the snRNA-seq data labeled by cell type. We clearly see the main two trajectories for germ cell and somatic cell differentiations and some other small clusters related to immune or erythroid cells. B UMAP of the snRNA-seq data where nuclei are colored in gray if they were called by both CellRanger-arc and EmptyDropsMultiome, blue if they were called by CellRanger-arc but not EmptyDropsMultiome, and red if they were called by EmptyDropsMultiome but not CellRanger-arc. CellRanger-arc is prone to missing nuclei in the germ cell differentiation trajectory due to the chromatin being more actively remodeled than in somatic cells [20]. C Bar plots showing the fraction of nuclei per cell type called by CellRanger-arc, EmptyDropsMultiome, or both. Oocytes are essentially missed by CellRanger-arc, which recovers only 17% of the ones EmptyDropsMultiome does. Abbreviations: ooc, oocytes; pre-ooc-2, late pre-oocytes; pre-ooc-1, early pre-oocytes; oog-m, oogonia-meiotic; oog-st, oogonia STRA8; PGC, primordial germ cells; CoEp, coelomic epithelium; supp1/2/3, early/middle/late supporting cells; Mes, mesenchymal; Imm, immune; Endo, endothelial; Ery, erythroid