scDetect: a rank-based ensemble learning algorithm for cell type identification of single-cell RNA sequencing in cancer

• DOI: 10.1093/bioinformatics/btab410
• Publication Date: 2021-05-27T12:46:35Z
• Authors: Yifei Shen; Qinjie Chu; Michael P Timko; Longjiang Fan
• Keywords: Ensemble Learning; Identification; Cell type; Bioconductor; Data type; Profiling (computer programming)

Abstract Motivation Single-cell RNA sequencing (scRNA-seq) has enabled the characterization of different cell types in many tissues and tumor samples. Cell type identification is essential for single-cell profiling, currently transforming life sciences. Often, this achieved by searching combinations genes that have previously been implicated as being cell-type specific, an approach not quantitative does explicitly take advantage other scRNA-seq studies. Batch effects data platforms greatly decrease predictive performance inter-laboratory validation. Results Here, we present a new ensemble learning method named ‘scDetect’ combines gene expression rank-based analysis majority vote machine-learning probability-based prediction capable highly accurate classification cells based on platforms. Because heterogeneity, order to accurately predict RNA-seq data, also incorporated copy number variation consensus clustering epithelial score classification. We applied scDetect from pancreatic tissue, mononuclear biopsies show classified individual with high accuracy better than publicly available tools. Availability implementation open source software. Source code test freely Github (https://github.com/IVDgenomicslab/scDetect/) Zenodo (https://zenodo.org/record/4764132#.YKCOlrH5AYN). The examples tutorial page at https://ivdgenomicslab.github.io/scDetect-Introduction/. And will be Bioconductor. Supplementary information are Bioinformatics online.