Detecting novel cell type in single-cell chromatin accessibility data via open-set domain adaptation

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Single-cell technologies generate vast multi-omics data, necessitating robust cell type annotation.
• A key challenge in single-cell data analysis is annotating unknown cell types not present in training datasets.
• Existing scATAC-seq annotation methods often fail to identify these unknown cell types.

Purpose of the Study:

• To develop an automated method for cell type annotation in scATAC-seq data that can detect unknown cell types.
• To address the limitations of current methods that only assign cells to known categories.

Main Methods:

• Developed OVAAnno, an automated cell type annotation method.
• Utilized open-set domain adaptation techniques to distinguish between known and unknown cell types.
• Applied the method to scATAC-seq and scRNA-seq datasets.

Main Results:

• OVAAnno successfully identified both known and unknown cell types in scATAC-seq data.
• The method demonstrated robust performance in detecting novel cell populations.
• OVAAnno also showed effectiveness when applied to single-cell RNA sequencing (scRNA-seq) data.

Conclusions:

• OVAAnno offers a significant advancement in single-cell data analysis by enabling the detection of unknown cell types.
• The method enhances the accuracy and completeness of cell type annotation for scATAC-seq and scRNA-seq data.
• The availability of the code facilitates further research and application in the field.