MiMeJF: Application of Coupled Matrix and Tensor Factorization (CMTF) for Enhanced Microbiome-Metabolome Multi-Omic...

Area of Science:

• Multi-omics integration
• Systems biology
• Biomarker discovery

Background:

• Microbiome and metabolome data integration offers profound biological insights.
• Stepwise multi-omic analysis limits detection accuracy and data potential.
• Synergistic analysis is crucial for a comprehensive understanding.

Purpose of the Study:

• To introduce a Coupled Matrix and Tensor Factorization (CMTF) framework for joint microbiome and metabolome data analysis.
• To overcome limitations of traditional stepwise multi-omic approaches.
• To facilitate dimension reduction and biomarker identification.

Main Methods:

• Developed two CMTF frameworks to factorize microbial taxa, functional pathways, and metabolites.
• Applied CMTF to diverse datasets including built environments and human gut samples (IBD, COVID-19).
• Validated the framework's performance in dimension reduction and biomarker detection.

Main Results:

• Identified biologically meaningful biomarkers: Bacteroides vulgatus and acylcarnitines for IBD; pyroglutamic acid and p-cresol for COVID-19.
• CMTF consistently outperformed traditional methods in dimension reduction and biomarker detection.
• Demonstrated the framework's robustness in uncovering biologically relevant insights.

Conclusions:

• The CMTF framework represents a significant advancement in multi-omics integration and analysis.
• Despite data requirements, it offers a powerful tool for biomedical, environmental, and agricultural research.
• Provides new avenues for research by revealing key microbial and metabolic associations.