Single-cell RNA sequencing (scRNA-seq) experiments provide opportunities to peer into complex tissues at single-cell resolution. However, insightful biological interpretation of scRNA-seq data relies upon precise identification of cell types. The ability to identify the origin of a cell quickly and accurately will greatly improve downstream analyses. We present Sargent, a transformation-free, cluster-free, single-cell annotation algorithm for rapidly identifying the cell types of origin based on cell type-specific markers. We demonstrate Sargent's high accuracy by annotating simulated datasets. Further, we compare Sargent performance against expert-annotated scRNA-seq data from human organs including PBMC, heart, kidney, and lung. We demonstrate that Sargent retains both the flexibility and biological interpretability of cluster-based manual annotation. Additionally, the automation eliminates the labor intensive and potentially biased user annotation, producing robust, reproducible, and scalable outputs. • Sargent is a transformation-free, cluster-free, single-cell annotation algorithm for rapidly identifying the cell types of origin based on cell type-specific markers. • Sargent retains both the flexibility and biological interpretability of cluster-based manual annotation. • Automation eliminates the labor intensive and potentially biased user annotation, producing robust, reproducible, and scalable outputs.
Full Text Finder