Screening Potential Drugs for COVID-19 Based on Bound Nuclear Norm Regularization.

Area of Science:

• Computational Biology
• Drug Discovery
• Bioinformatics

Background:

• The COVID-19 pandemic caused by SARS-CoV-2 necessitates urgent development of effective therapeutic strategies.
• Identifying novel drug candidates against SARS-CoV-2 is crucial for global health.
• Existing methods for predicting antiviral drugs require enhancement for accuracy and efficiency.

Purpose of the Study:

• To develop and validate a novel computational method for predicting anti-SARS-CoV-2 drug candidates.
• To identify potential therapeutic agents for COVID-19 treatment.
• To investigate the molecular interactions of potential drugs with SARS-CoV-2 targets.

Main Methods:

• Construction of a heterogeneous virus-drug network using three association datasets.
• Integration of genomic sequences and chemical structures with Gaussian association profiles to compute virus and drug similarities.
• Application of a Bounded Nuclear Norm Regularization (BNNR) model, termed VDA-GBNNR, for drug candidate prediction and validation via fivefold cross-validation.

Main Results:

• The VDA-GBNNR model achieved superior performance with Area Under the Curve (AUC) values of 0.8965, 0.8562, and 0.8803 on three independent datasets.
• Six potential anti-SARS-CoV-2 drugs were identified: remdesivir, favipiravir, ribavirin, mycophenolic acid, niclosamide, and mizoribine.
• Molecular docking revealed significant binding energies for niclosamide (-8.06 kcal/mol) and mizoribine (-7.06 kcal/mol) with the SARS-CoV-2 spike protein and ACE2 interface, highlighting G496 and K353 as key residues.

Conclusions:

• The VDA-GBNNR method demonstrates high accuracy in predicting anti-SARS-CoV-2 drug candidates.
• Niclosamide and mizoribine show promising therapeutic potential for COVID-19 treatment.
• The findings provide valuable insights for the development of novel antiviral therapies against SARS-CoV-2.