Fig. 8
From: Quantile normalization of single-cell RNA-seq read counts without unique molecular identifiers
Quasi-UMI normalization improves resolution of pancreatic endocrine cells in the non-UMI Segerstolpe dataset. Each panel represents a marker gene for a specific cell type. Box plots show gene expression levels for cells that were annotated to each cell type by the original authors. Open circles indicate gene expression levels for cells that were unable to be annotated by the original authors. These cells were unambiguously assigned based on a random forest classifier applied to a GLM-PCA dimension reduction of QUMI count values. The overlap between annotated and reclassified cells validates the accuracy of the classification