Ensemble learning for classifying single-cell data and projection across reference atlases

Single-cell data are being generated at an accelerating pace. How best to project data across single-cell atlases is an open problem. We developed a boosted learner that overcomes the greatest challenge with status quo classifiers: low sensitivity, especially when dealing with rare cell types. By comparing novel and published data from distinct scRNA-seq modalities that were acquired from the same tissues, we show that this approach preserves cell-type labels when mapping across diverse platforms. We developed an ensemble classifier of scRNA-seq, single-nuclei RNA-sequencing (snRNA-seq), and bulk-extraction RNA-sequencing (RNA-seq) data: Ensemble Learning for classifying Single-cell data and projection across reference Atlases (ELSA; https://github.com/diazlab/ELSA). We trained ELSA on public atlases and tested it on published single-cell data, novel scRNA-seq and snRNA-seq of human glioma tissues (4 patients, >11K cells, Table S1 and S2). Note: Submitter did not submit the raw data files due to privacy concerns for patients.