A NMF based approach for integrating multiple data sources to predict HIV-1-human PPIs

Area of Science:

• Computational Biology
• Bioinformatics
• Systems Biology

Background:

• Predicting novel interactions between HIV-1 and human proteins is crucial for HIV research.
• Current prediction methods often rely on single biological data sources.

Purpose of the Study:

• To develop a novel framework for predicting protein-protein interactions (PPIs) between HIV-1 and human proteins.
• To integrate multiple biological data sources for enhanced prediction accuracy.

Main Methods:

• Developed a framework integrating multiple biological data sources via Non-negative Matrix Factorization (NMF).
• Converted data sets to biological networks, predicted modules, and combined them into meta-modules using NMF-based clustering.
• Analyzed Gene Ontology (GO) terms, KEGG pathways, and topological properties of human proteins in meta-modules.
• Performed statistical significance tests to validate predictions.

Main Results:

• Successfully predicted novel HIV-1 and human protein-protein interactions.
• Identified significant GO terms and KEGG pathways associated with predicted interactions.
• Human proteins in predicted interactions exhibit 'hub' and 'bottleneck' characteristics.
• Statistical significance tests confirmed the validity of the predictions.

Conclusions:

• The proposed NMF-based approach effectively integrates diverse biological data for predicting HIV-1-human PPIs.
• Predicted interactions are largely supported by existing literature.
• The findings highlight the statistical significance and biological relevance of the predicted interactions.