Indicator Regularized Non-Negative Matrix Factorization Method-Based Drug Repurposing for COVID-19.

Area of Science:

• Virology
• Computational Biology
• Drug Discovery

Background:

• COVID-19 is a severe global infectious disease with limited treatment options.
• Drug repurposing offers a faster and more cost-effective approach to finding new therapies compared to de novo drug discovery.
• Existing research highlights the need for efficient methods to identify potential virus-drug associations.

Purpose of the Study:

• To develop and validate a novel computational method for predicting virus-drug associations.
• To assess the efficacy of the proposed method in prioritizing potential drug candidates for viral infections, including COVID-19.
• To contribute to the acceleration of drug discovery for emerging infectious diseases.

Main Methods:

• Construction of a comprehensive virus-drug dataset comprising 34 viruses, 210 drugs, and 437 confirmed virus-drug pairs.
• Development of the Indicator Regularized non-negative Matrix Factorization (IRNMF) algorithm, incorporating indicator matrices and Karush-Kuhn-Tucker conditions.
• Performance evaluation using 5-fold cross-validation, comparing IRNMF against existing methods.

Main Results:

• The IRNMF method demonstrated superior performance compared to other evaluated methods.
• The Area Under the receiver operating characteristic Curve (AUC) for IRNMF reached 0.8127, indicating high predictive accuracy.
• Analysis of the COVID-19 case confirmed the algorithm's ability to prioritize previously unknown virus-drug associations.

Conclusions:

• IRNMF is a robust and effective computational approach for predicting virus-drug associations.
• The developed method shows significant potential for accelerating drug repurposing efforts, particularly for novel infectious diseases like COVID-19.
• This study provides a valuable tool for identifying promising drug candidates to combat viral outbreaks.