scClassifR: Framework to accurately classify cell types in single-cell RNA-sequencing data

Abstract

MotivationAutomatic cell type identification in scRNA-seq datasets is an essential method to alleviate a key bottleneck in scRNA-seq data analysis. While most existing tools show good sensitivity and specificity in classifying cell types, they often fail to adequately not-classify cells that are not present in the used reference. ResultsscClassifR is a novel R package that provides a complete framework to automatically classify cells in scRNA-seq datasets. It supports both Seurat and Bioconductors SingleCellExperiment and is thereby compatible with the vast majority of R-based analysis workflows. scClassifR uses hierarchically organised SVMs to distinguish a specific cell type versus all others. It shows comparable or even superior sensitivity and specificity compared to existing tools while being robust in not-classifying unknown cell types. As a unique feature, it reports ambiguous cell assignments, including the respective probabilities. Finally, scClassifR provides dedicated functions to train and evaluate classifiers for additional cell types. Availability and ImplementationscClassifR is freely available on GitHub (https://github.com/grisslab/scClassifR).