A Generalized Higher-order Correlation Analysis Framework for Multi-Omics Network Inference.

Area of Science:

• Computational biology
• Systems biology
• Bioinformatics

Background:

• Multi-omics data integration is crucial for understanding complex biological systems and diseases.
• Existing methods struggle with high dimensionality, higher-order correlations, and flexibility in analyzing omics-to-phenotype relationships.

Purpose of the Study:

• To introduce a novel pipeline, Sparse Generalized Tensor Canonical Correlation Analysis Network Inference (SGTCCA-Net), for robust multi-omics network construction.
• To address limitations of existing canonical correlation methods in handling complex, high-dimensional omics data.

Main Methods:

• Developed SGTCCA-Net, a novel pipeline for multi-omics network analysis.
• Implemented tensor canonical correlation analysis to capture higher-order correlations.
• Incorporated sparsity and flexibility for focused correlation analysis (omics-to-omics and omics-to-phenotype).

Main Results:

• SGTCCA-Net effectively overcomes limitations of previous methods for multi-omics integration.
• The pipeline demonstrates computational efficiency and flexibility in network inference.
• Simulation and real-data experiments validate the method's effectiveness in identifying key omics networks and features.

Conclusions:

• SGTCCA-Net provides a powerful and flexible approach for multi-omics network inference.
• The method enhances the understanding of relationships between molecular features across different omics layers.
• This pipeline facilitates downstream analyses and biological discovery from complex datasets.