scAnno: a deconvolution strategy-based automatic cell type annotation tool for single-cell RNA-sequencing data sets.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) provides crucial insights into cellular heterogeneity.
• Accurate cell type annotation is essential for analyzing complex scRNA-seq datasets.
• Existing annotation tools may lack robustness and compatibility with diverse datasets.

Purpose of the Study:

• To develop an automated, robust, and accurate tool for scRNA-seq data annotation.
• To introduce a novel methodology based on joint deconvolution and logistic regression for cell typing.
• To identify cell type-specific marker genes from scRNA-seq data.

Main Methods:

• Development of scAnno, an automated annotation tool for scRNA-seq data.
• Construction of comprehensive reference profiles for human and mouse cell types and tissues.
• Utilizing a joint deconvolution strategy and logistic regression for cell type annotation.
• Employing co-expression and seed genes to identify cell type-specific marker genes.

Main Results:

• scAnno achieved high accuracy in cell type annotation (99.05% internally, 95.56% cross-platform).
• The tool accurately identified cell type-specific marker genes, validated against the CellMarker database.
• scAnno demonstrated superior performance compared to existing annotation tools like SingleR, scPred, CHETAH, and scmap-cluster.

Conclusions:

• scAnno represents a novel methodology for automated cell typing in scRNA-seq analysis.
• The tool offers flexibility, interpretability, and superior accuracy in cell type annotation and marker gene identification.
• scAnno is a valuable application for broader scRNA-seq data analysis and discovery.