scICML: Information-Theoretic Co-Clustering-Based Multi-View Learning for the Integrative Analysis of Single-Cell...

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• High-throughput sequencing enables multi-omics profiling of single cells, revealing cellular heterogeneity.
• Single-cell multi-omics data present challenges due to high noise and sparsity.
• Integrating diverse molecular data layers is crucial for comprehensive cellular analysis.

Purpose of the Study:

• To develop a novel method for effective multi-omics single-cell data integration.
• To address the challenges of noise and sparsity in single-cell multi-omics datasets.
• To improve the accuracy and biological interpretability of single-cell data analysis.

Main Methods:

• Developed single-cell information-theoretic co-clustering-based multi-view learning (scICML).
• Utilized co-clusterings to aggregate similar features within each data modality.
• Implemented automatic matching of linked features across different data types to capture biological dependencies.

Main Results:

• scICML successfully integrated multi-omics single-cell data from four real-world datasets.
• The method demonstrated improved overall clustering performance compared to existing approaches.
• Analysis of peripheral blood mononuclear cells yielded significant biological insights.

Conclusions:

• scICML provides a robust framework for multi-omics single-cell data integration.
• The method effectively handles noise and sparsity, uncovering common cellular patterns.
• scICML enhances the biological understanding derived from complex single-cell datasets.