Machine Learning and Deep Learning Frameworks for Human-Virus Protein-Protein Interaction Prediction: Emerging...

Area of Science:

• Virology
• Computational Biology
• Machine Learning

Background:

• Coronavirus disease 2019 (COVID-19), caused by SARS-CoV-2, is a major global health crisis.
• SARS-CoV-2, a beta-coronavirus, exhibits high transmissibility and causes severe illness.
• Understanding virus-host protein-protein interactions (PPIs) is crucial for identifying therapeutic targets.

Purpose of the Study:

• To review computational approaches for predicting viral protein-protein interactions (PPIs).
• To highlight machine learning (ML) and deep learning (DL) techniques for PPI prediction.
• To discuss the generalizability of models across different viral families.

Main Methods:

• Review of existing literature on computational PPI prediction methods.
• Analysis of ML and DL techniques utilizing diverse biological data (sequences, structures, genomics, etc.).
• Evaluation of performance metrics and limitations in benchmark comparability.

Main Results:

• Computational methods, particularly ML/DL, provide efficient and scalable solutions for PPI prediction.
• Diverse biological data sources enhance the accuracy of PPI prediction models.
• Challenges remain in standardizing evaluation practices and ensuring model generalizability.

Conclusions:

• ML/DL-based computational approaches are vital for advancing the study of viral pathogenesis and drug discovery.
• Further research is needed to refine prediction models and improve their applicability across various viral families.
• Standardized benchmarks are essential for comparing the performance of different PPI prediction methods.