Combination therapy synergism prediction for virus treatment using machine learning models.

Area of Science:

• Virology
• Computational Biology
• Pharmacology

Background:

• Synergistic drug combinations are crucial for effective treatments.
• Limited research exists on computational prediction of antiviral combination therapies.
• Developing novel antiviral strategies is a global health priority.

Purpose of the Study:

• To propose and evaluate AI-based models for predicting synergistic antiviral drug combinations.
• To establish the first dataset and machine learning model for viral combination therapy prediction.
• To identify novel synergistic antiviral drug combinations for various viral diseases.

Main Methods:

• Assembled a comprehensive dataset of viral strains, drug compounds, and interactions.
• Developed and trained machine learning models including Random Forest, Support Vector Machine (SVM), and deep learning models.
• Validated model predictions using statistical tests (t-test).

Main Results:

• Machine learning models demonstrated high performance in predicting synergistic antiviral combinations.
• The study presents the first dataset and learning model specifically for viral combination therapy.
• One predicted combination (acyclovir and ribavirin) was experimentally validated for synergistic antiviral activity against herpes simplex type-1 virus.

Conclusions:

• AI-based models are effective for predicting synergistic antiviral drug combinations.
• The developed models and dataset represent a significant advancement in antiviral drug discovery.
• This approach holds promise for accelerating the development of novel treatments for viral infections.