Multi-view Subspace Clustering Analysis for Aggregating Multiple Heterogeneous Omics Data.

Area of Science:

• Computational Biology
• Bioinformatics
• Data Integration

Background:

• Integrating distinct biological data types is crucial for understanding complex biological processes and diseases.
• Identifying intrinsic data structures within multiple embedded subspaces presents a significant challenge for conventional integration methods.

Purpose of the Study:

• To propose a novel framework, Multi-view Subspace Clustering Analysis (MSCA), for effective biological data integration.
• To capture underlying sample heterogeneity by measuring local similarities and obtaining global consensus patterns across multiple data types.

Main Methods:

• Developed the Multi-view Subspace Clustering Analysis (MSCA) framework.
• Applied MSCA to synthetic datasets to evaluate its pattern recognition and noise robustness.
• Utilized MSCA on the Cancer Cell Line Encyclopedia (CCLE) dataset for real-world biological data analysis.

Main Results:

• MSCA effectively recognized predefined sample patterns in synthetic datasets and demonstrated robustness to noise.
• MSCA successfully identified cell clusters with common aberrations across cancer types in the CCLE dataset.
• MSCA exhibited superior performance compared to state-of-the-art methods like iClusterPlus, SNF, and ANF in simulations and case studies.

Conclusions:

• MSCA provides a comprehensive approach to capturing sample heterogeneity from multi-view biological data.
• The framework offers significant advantages over existing methods for biological data integration and analysis.
• MSCA holds promise for advancing our understanding of complex diseases through integrated omics data.