Platinum(IV) compounds as potential drugs: a quantitative structure-activity relationship study.

Area of Science:

• Computational chemistry
• Medicinal chemistry
• Drug discovery

Background:

• Machine learning (ML) and increased computing power are revolutionizing drug design and repurposing.
• Predictive models can accelerate the identification of novel therapeutic agents.

Purpose of the Study:

• To predict the bioactivity of novel platinum(IV) complexes using ML.
• To develop and validate Quantitative Structure-Activity Relationship (QSAR) models for predicting activities against SARS-CoV.

Main Methods:

• Utilized ML predictive models available at chemosophia.com.
• Developed and validated two novel QSAR models based on the BiS algorithm.
• Employed 10-fold cross-validation to assess model predictive power (cross-R² 0.863–0.903).

Main Results:

• Predicted 38 diverse activities for eighteen platinum(IV) complexes.
• Activities spanned antioxidant, antibacterial, antiviral, anti-inflammatory, anti-arrhythmic, and anti-malarial potentials.
• Identified complexes 1, 3, and 13 as potential inhibitors of SARS-CoV RNA-dependent RNA polymerase.

Conclusions:

• The developed QSAR models demonstrate high predictive accuracy.
• Platinum(IV) complexes 1, 3, and 13 are promising candidates for SARS-CoV inhibition.
• ML-driven prediction facilitates efficient drug discovery and repurposing efforts.