Fig. 5
From: The Poisson distribution model fits UMI-based single-cell RNA-sequencing data
The t-SNE (a, c) and UMAP (b, d) visualizations of the A5 sample which consists of n=1,476 cells [28]. The top two panels (a, b) were based on Hclust-Departure using model departure as data representation. The bottom two panels (c, d) were labeled by cell types from the Seurat analysis of Cheng et al. [28]. The clusters discovered by Hclust-Departure are consistent with those identified by Seurat. Furthermore, Hclust-Departure identifies several significant subclusters (namely B-cells, NK cells and erythrocytes)