scAGCI: an anchor graph-based method for cell clustering from integrated scRNA-seq and scATAC-seq data.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Single-cell multi-omics clustering faces challenges from data noise and heterogeneity.
• Existing anchor graph methods struggle to model higher-order feature interactions effectively.

Purpose of the Study:

• To develop a novel cell clustering method, scAGCI, that integrates single-cell RNA sequencing (scRNA-seq) and single-cell Assay for Transposase-Accessible Chromatin sequencing (scATAC-seq) data.
• To address limitations in modeling higher-order feature interactions in current multi-view anchor graph approaches.

Main Methods:

• scAGCI utilizes specific and shared anchor graphs to represent omics data properties during dynamic anchor unification.
• The method mines high-order shared information to enhance omics representation.
• Clustering is achieved by integrating these specific and shared omics representations.

Main Results:

• scAGCI demonstrated superior clustering performance and computational efficiency compared to 13 state-of-the-art methods.
• The method excels in cell-type identification and subtype resolution.
• Benchmarking confirmed the preservation of biologically meaningful omics patterns through marker gene enrichment and functional analyses.

Conclusions:

• scAGCI is a robust tool for elucidating cellular heterogeneity in single-cell multi-omics data.
• The method effectively integrates scRNA-seq and scATAC-seq data, overcoming limitations of previous approaches.