scATAnno: Automated Cell Type Annotation for single-cell ATAC Sequencing Data.

Area of Science:

• Genomics
• Computational Biology
• Epigenetics

Background:

• Single-cell epigenomic techniques like scATAC-seq are advancing rapidly.
• Accurate cell type identification from scATAC-seq data is crucial for biological interpretation.
• Existing methods may require complementary data or lack robustness.

Purpose of the Study:

• To introduce scATAnno, a Python package for automated scATAC-seq data annotation.
• To enable cell type identification using large-scale scATAC-seq reference atlases.
• To provide a robust tool for interpreting complex biological systems via epigenomic data.

Main Methods:

• Developed scATAnno, a Python package for scATAC-seq analysis.
• Generated reference atlases from public scATAC-seq datasets.
• Integrated query data with reference atlases for annotation, avoiding scRNA-seq data.
• Incorporated KNN-based and weighted distance-based uncertainty scores for enhanced accuracy.

Main Results:

• scATAnno demonstrated superior performance compared to 7 other cell annotation approaches across multiple datasets and metrics.
• Successfully annotated cell types in diverse datasets including PBMCs, TNBC, and BCC.
• Effectively detected distinct cell populations not present in reference atlases.
• Validated utility across various biological conditions.

Conclusions:

• scATAnno is a powerful tool for scATAC-seq reference building and cell type annotation.
• It facilitates accurate interpretation of new scATAC-seq datasets.
• The package enhances understanding of cellular heterogeneity in complex biological systems.