Estimation of the Binding Affinities of Glycogen Phosphorylase Inhibitors by Molecular Docking to Support the Treatment of Type 2 Diabetes

In the current landscape of drug discovery, various docking programs for virtual database screening significantly reduce costs and time. This study re-docked twenty-three known inhibitors of glycogen phosphorylase (GP), a key target for type 2 diabetes (T2D) therapy, using seven methods including Autodock 4 (AD4), AutoDockVina (Vina), modified Vina (mVina), Standard Precision mode (SP) and Extra Precision mode (XP) of Glide methods, Molecular Operating Environment (MOE) and Genetic Optimization for Ligand Docking (GOLD). Results showed that GOLD showed the worst computational precision with the highest RMSE of 20.98 kcal mol(-1). Conversely, MOE was the most precise with the lowest RMSE of 1.99 kcal mol(-1), closely followed by AD4 (2.27 kcal mol(-1)). However, MOE failed to generate the correct ligand-binding pose, showing a 0% success rate in docking for all RMSD resolutions (<0.2, 0.15, and 0.1 nm). Among the top-performing methods, GOLD surpassed others in docking success rates for GP ligands, achieving 96% success at RMSD < 0.2 nm, compared to 74%, 70%, and 74% for AD4, Vina, and mVina, respectively. These four packages can produce a ligand-binding posture that closely resembles the crystal structure discovered through experimental studies. The findings serve as the foundation for selecting an appropriate tool for screening candidate drugs for the T2D therapeutic target.