PT  - JOURNAL ARTICLE
AU  - Aleksandr Ianevski
AU  - Anil K Giri
AU  - Tero Aittokallio
TI  - Fully-automated cell-type identification with specific markers extracted from single-cell transcriptomic data
AID  - 10.1101/812131
DP  - 2019 Jan 01
TA  - bioRxiv
PG  - 812131
4099  - http://biorxiv.org/content/early/2019/10/21/812131.short
4100  - http://biorxiv.org/content/early/2019/10/21/812131.full
AB  - Single-cell transcriptomics enables systematic charting of cellular composition of complex tissues. Identification of cell populations often relies on unsupervised clustering of cells based on the similarity of the scRNA-seq profiles, followed by manual annotation of cell clusters using established marker genes. However, manual selection of marker genes for cell-type annotation is a laborious and error-prone task since the selected markers must be specific both to the individual cell clusters and various cell types. Here, we developed a computational method, termed ScType, which enables data-driven selection of marker genes based solely on given scRNA-seq data. Using a compendium of 7 scRNA-seq datasets from various human and mouse tissues, we demonstrate how ScType enables unbiased, accurate and fully-automated single-cell type annotation by guaranteeing the specificity of marker genes both across cell clusters and cell types. The widely-applicable method is implemented as an interactive web-tool (https://sctype.fimm.fi), connected with comprehensive database of specific markers.