scATAcat: cell-type annotation for scATAC-seq data

Area of Science:

• Genomics
• Computational Biology
• Single-cell analysis

Background:

• Single-cell ATAC sequencing (scATAC-seq) profiles the chromatin accessibility landscape of individual cells.
• Annotating cell types in scATAC-seq data is difficult due to the lack of established marker regions, unlike in scRNA-seq.
• Existing methods often translate scATAC-seq data to expression space, relying on gene expression patterns for annotation.

Purpose of the Study:

• To develop a novel computational approach for cell-type annotation directly from scATAC-seq data.
• To leverage characterized bulk ATAC-seq data as reference prototypes for annotation.
• To address the challenges of sparsity and lack of marker regions in scATAC-seq data.

Main Methods:

• Proposed scATAcat, a novel method for annotating scATAC-seq data.
• Utilized characterized bulk ATAC-seq data as prototypes.
• Aggregated cells within clusters to create pseudobulk data, mitigating single-cell data sparsity.

Main Results:

• Demonstrated the feasibility and performance of scATAcat using multiple annotated datasets.
• Quantified the accuracy of cell-type annotations generated by scATAcat.
• scATAcat provides a viable alternative for cell-type identification in scATAC-seq.

Conclusions:

• scATAcat offers an effective strategy for cell-type annotation in scATAC-seq data.
• The method successfully utilizes bulk ATAC-seq prototypes and pseudobulking to overcome data limitations.
• scATAcat is available as a Python package, facilitating its adoption in the research community.