Modelling of AgrA inhibitors to combat anti-microbial resistance in .

is a Gram-positive bacterium found on human skin that causes skin and soft tissue infections, as well as pneumonia, osteomyelitis, and endocarditis. The prevalence of antibiotic resistant strains has made the treatments less effective. An efficient alternate method for battling these contagious diseases is anti-virulence strategy. The AgrA protein, a key activator of Accessory Gene Regulator system in S. aureus, is vital to the virulence of the organism and, consequently, its pathogenesis. Using a Machine Learning algorithm, the Support Vector Machine (SVM), and a ligand-based pharmacophore modelling method, prediction models of AgrA inhibitors were developed. The metrics of the SVM model were inadequate, hence it was not used for virtual screening. For ligand-based pharmacophore modelling, 14 of 29 compounds were removed from the active set due to a lack of shared pharmacophore properties, and 504 compounds were designated as decoys. A 3D pharmacophore model was created using LigandScout 4.4.5, with a fit score of 57.48, including a positive ionizable group, one hydrogen bond donor, and three hydrogen bond acceptors. The model after further validation was used to virtually screen an external database which resulted in six hits. These compounds were docked with the AgrA domain crystal structure to determine the inhibitor activity. Further, each docked complex was subjected to a 100 ns molecular dynamics simulation. CID238 and CID20510252 demonstrated potent inhibitory binding interactions and hence can be used to develop AgrA inhibitors in future after proper validation.Communicated by Ramaswamy H. Sarma.