Modeling Antimicrobial Activity of Cannabinoid Derivatives Against MRSA: A QSAR Approach Using Machine Learning

The objective of this research is to establish a dependable correlation Methicillin-resistant Staphylococcus aureus (MRSA) infections are becoming more common, which is a serious public health concern. This work investigates the potential of synthetic and natural phytocannabinoids as MRSA inhibitors. Five predictive models were created using 2D molecular descriptors and QSAR analysis, ADMET prediction, and molecular docking. These models are MLR, SVR, LASSO, Ridge, and polynomial regression. Out of all of these, the polynomial regression model performed the best, exhibiting excellent reliability in predicting anti-MRSA activity with an R2 of 99.86% and an RMSE of 2.29. The design of novel cannabinoid-based MRSA inhibitors is guided by machine learning and molecular descriptor selection, which is highlighted by this method.