RESUMO
Introduction: Machine learning methods, coupled with a tremendous increase in computer power in recent years, are promising tools in modern drug design and drug repurposing. Methods: Machine learning predictive models, publicly available at chemosophia.com, were used to predict the bioactivity of recently synthesized platinum(IV) complexes against different kinds of diseases and medical conditions. Two novel QSAR models based on the BiS algorithm are developed and validated, capable to predict activities against the SARS-CoV virus and its RNA dependent RNA polymerase. Results: The internal predictive power of the QSAR models was tested by 10-fold cross-validation, giving cross-R2 from 0.863 to 0.903. 38 different activities, ranging from antioxidant, antibacterial, and antiviral activities, to potential anti-inflammatory, anti-arrhythmic and anti-malarial activity were predicted for a series of eighteen platinum(IV) complexes. Conclusion: Complexes 1, 3 and 13 have high generalized optimality criteria and are predicted as potential SARS-CoV RNA dependent RNA polymerase inhibitors.