 This study used computational docking techniques to identify potential drug targets for 296 essential proteins and 218 antibiotics. The results showed that many proteins were able to bind multiple drugs, suggesting that these drugs may be effective against a wide range of bacteria. Additionally, the study found that rescoring docking poses using machine learning-based approaches improved model performance, resulting in average area under receiver operating characteristic, AUROC, scores of up to 0.63. These results suggest that further improvements in modeling protein ligand interactions could help to better utilize Alpha Fold 2 for drug discovery. This article was authored by Felix Wong, Arthakrishnan, Erika Jayjing and others.