Integrating molecular docking and chemoinformatics for predicting antimicrobial activity of essential oils

Essential oils (EOs) are complex mixtures of natural metabolites known for their antimicrobial properties, yet their molecular mechanisms of action remain poorly understood. This study addresses the need for predictive tools to identify bioactive EO constituents by integrating molecular docking and chemoinformatics approaches. A comprehensive ligand library comprising 7,222 EO-derived compounds was virtually screened against a panel of microbial protein targets using a calibrated docking workflow supported by five distinct scoring functions. Agglomerative hierarchical clustering of high-scoring ligands, combined with the Euclidean Proximity Normalized Matrix to Cheminformatics Space (EPNMCS) metric, enabled prediction of target-specific protein–metabolite interactions. Experimental validation via in vitro antimicrobial assays demonstrated that Mentha piperita exhibited the strongest inhibitory activity, whereas Chamaemelum nobile displayed the widest predicted interaction spectrum. Notably, α-bisabolol and its oxidized derivatives showed high binding affinity for key microbial proteins, particularly FtsZ and Erg11. Despite a low correlation between computational predictions and minimum inhibitory concentration (MIC) values, the integrative strategy successfully identified EO metabolites with potential antimicrobial activity. These findings underscore the value of combining in silico and in vitro approaches to elucidate EO-target interactions and guide the discovery of natural antimicrobial agents.