To accelerate the discovery of novel efflux pump modulators against multidrug-resistant Gram-negative bacteria, we have developed a predictive chem-bioinformatics platform. This tool translates chemical structures (SMILES) into a rich descriptor space, subsequently processed by pre-validated XGBoost machine learning models. It provides rapid in silico predictions of crucial efficacy parameters: Minimum Inhibitory Concentration (MIC) and comparative docking scores for the clinically significant RND-type efflux pumps AcrB and MexB. By enabling swift profiling and prioritization of new chemical entities, and benchmarking them against known agents, this platform serves as a practical engine for hit generation, complementing structural and mechanistic investigations into efflux pump assembly, function, and inhibition.
Developed by Sadegh Zargan, Dr. Ben Luisi.
(Initial version; under testing and review.)
For more information: Zargansadegh@gmail.com