Realizing drug repositioning by adapting a recommendation system to handle the process

Area of Science:

• Computational drug discovery
• Pharmacology
• Bioinformatics

Background:

• Drug repositioning identifies new therapeutic indications for existing drugs.
• This approach mitigates risks, costs, and timelines associated with traditional drug discovery.
• Computational methods integrating drug and disease features are increasingly utilized.

Purpose of the Study:

• To develop an enhanced computational method for drug repositioning.
• To predict new indications for target drugs by integrating diverse feature sets.
• To frame drug repositioning as a recommendation problem for a novel perspective.

Main Methods:

• Combined multiple features: chemical structures, protein interactions, and side-effects.
• Employed a recommendation system based on Pareto dominance and collaborative filtering.
• Integrated multiple data sources and features for prediction.

Main Results:

• The proposed method achieved concentrated predictions with high precision.
• Nearly half of the method's predictions were accurate.
• Performance was evaluated across several computational settings.

Conclusions:

• The developed recommendation method demonstrates high precision in drug repositioning.
• This approach offers superior predictive accuracy compared to existing state-of-the-art methods.
• The study validates the effectiveness of recommendation techniques for drug repositioning.