Determining human-coronavirus protein-protein interaction using machine intelligence.

Area of Science:

• Computational Biology
• Virology
• Machine Learning

Background:

• The Severe Acute Respiratory Syndrome CoronaVirus 2 (SARS-CoV-2) pandemic has caused millions of deaths globally.
• Protein-Protein Interactions (PPIs) between SARS-CoV-2 and human proteins are critical for viral pathogenesis but remain poorly understood.
• Investigating these host-viral PPIs is crucial for developing effective antiviral strategies.

Purpose of the Study:

• To elucidate host-viral PPIs using machine learning (ML) approaches.
• To identify potential SARS-CoV-2 human target proteins.
• To validate the biological significance of predicted interactions.

Main Methods:

• Developed ML classifiers using five sequence-based features of human proteins: Amino Acid Composition, Pseudo Amino Acid Composition, Conjoint Triad, Dipeptide Composition, and Normalized Auto Correlation.
• Employed an ensemble method combining Random Forest Model (RFM), AdaBoost, and Bagging with a majority voting rule.
• Validated predicted interactions using Gene Ontology (GO) and KEGG pathway enrichment analysis.

Main Results:

• The proposed ensemble ML model achieved encouraging statistical performance.
• Predicted 111 potential SARS-CoV-2 human target proteins with a high likelihood factor (≥70%).
• Validated the biological significance of these predicted interactions through GO and KEGG pathway analysis.

Conclusions:

• This study provides a deeper understanding of the molecular mechanisms underlying SARS-CoV-2 pathogenesis.
• Identified potential therapeutic targets for developing novel anti-COVID medications.
• Highlights the utility of ML in uncovering critical host-viral interactions.