Discovery of new chemical scaffolds for none-nucleoside reverse transcriptase inhibitors using chemoinformatic methods

Non-nucleoside reverse transcriptase inhibitors (NNRTIs) have gained a definitive place due to theiruniqueantiviral potency, high specificity and low toxicity in antiretroviralcombination therapies used totreat HIV. To design more specific HIV-1 inhibitors, 218 Non-nucleoside reverse transcriptaseinhibitorswith their EC50 values were collected from different literature sources. Then, different types offingerprint descriptors were calculated by PaDel descriptor and afterwards enhanced replacement method(ERM) was used as the variable selection approach to choose more relevant variables. Based on selecteddescriptors, a classification SVM model was constructed to categorize compounds into two groups ofactive and inactive.The most potent inhibitor available in the set was used to carry out a similarity searchin PubChem server to retrieve compounds similar to the hit compound. The screened compounds withabove 85 percent of similarity to the hit compound were used as the input to the SVM model. Likewise,the most active compound in the set has been used as the input to the Pharmit web server to screen theMolport library by constructing structure based pharmacophre models. Shape filters for protein and ligandand also Lipinski rule of five were used to filter the output of virtual screening from resultedpharmacophore models. 53 compounds from the structure based pharmacophore search and the activecompounds form the SVM model were docked into the protein with the PDB code of 3DLG to investigatethe important molecular interactions and five ligands showing appropriate interactions with the targetwere subjected to experimental studies.

کلیدواژه ها:

Similarity search ، Pharmacophore modeling ، SVM ، Docking