New Insights Into Drug Repurposing for COVID-19 Using Deep Learning

Area of Science:

• Pharmacology
• Artificial Intelligence
• Infectious Diseases

Background:

• The COVID-19 pandemic necessitates rapid identification of effective treatments.
• Drug repurposing is a key strategy to accelerate the availability of anti-COVID-19 medications.
• Existing deep learning (DL) methods for drug discovery often lack interpretability.

Purpose of the Study:

• To review current drug repurposing strategies for COVID-19.
• To propose enhancements to deep learning models for more effective drug repurposing.
• To transform DL models from "black boxes" to "glass boxes" with interpretable rationale.

Main Methods:

• Review of existing literature on drug repurposing for COVID-19.
• Proposal for an improved deep learning approach with optimized hyperparameters, activation functions, and optimization algorithms.
• Focus on feature extraction from big data for enhanced model performance.

Main Results:

• The improved DL approach aims for high prediction accuracy while providing interpretability.
• This approach can establish cause-and-effect relationships for drug suitability in COVID-19 treatment.
• The enhanced DL model is expected to outperform traditional "black box" DL models.

Conclusions:

• The proposed enhanced DL approach offers a new generation of tools for drug repurposing.
• This methodology can be extended to other complex diseases and the development of novel treatment strategies.
• The approach supports precision medicine and the discovery of new therapeutic agents for COVID-19 and beyond.