A universal framework for single-cell multi-omics data integration with graph convolutional networks.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Single-cell omics data generation is rapidly increasing.
• Integrating diverse single-cell omics datasets presents significant challenges due to variations in sequencing technologies, data quality, and expression patterns.

Purpose of the Study:

• To develop a universal framework for integrating single-cell multi-omics data.
• To address the limitations of current methods in handling heterogeneous single-cell data.

Main Methods:

• A novel framework, Graph Convolutional Network for Single-Cell (GCN-SC), was developed.
• GCN-SC utilizes a mutual nearest neighbor algorithm to identify cell-pairs and constructs a mixed graph.
• Graph convolutional networks adjust count matrices, followed by non-negative matrix factorization for dimension reduction and visualization.

Main Results:

• GCN-SC successfully integrated single-cell sequencing data from multiple technologies, species, and omics types.
• The framework demonstrated superior performance compared to state-of-the-art methods like Seurat, LIGER, GLUER, and Pamona.
• Evaluation across six diverse datasets confirmed the effectiveness and robustness of GCN-SC.

Conclusions:

• GCN-SC provides an effective and universal solution for single-cell multi-omics data integration.
• The proposed method enhances the ability to analyze complex biological systems by merging heterogeneous datasets.
• GCN-SC represents a significant advancement in single-cell data analysis, enabling more comprehensive biological discoveries.