Dolezal_2015_J.Mol.Graph.Model_56_113

To predict unknown reactivation potencies of 12 mono- and bis-pyridinium aldoximes for VX-inhibited rat acetylcholinesterase (rAChE), three-dimensional quantitative structure-activity relationship (3D QSAR) analysis has been carried out. Utilizing molecular interaction fields (MIFs) calculated by molecular mechanical (MMFF94) and quantum chemical (B3LYP/6-31G*) methods, two satisfactory ligand-based CoMFA models have been developed: 1. R2=0.9989, QLOO2=0.9090, QLTO2=0.8921, QLMO(20%)2=0.8853, Rext2=0.9259, SDEPext=6.8938; 2. R2=0.9962, QLOO2=0.9368, QLTO2=0.9298, QLMO(20%)2=0.9248, Rext2=0.8905, SDEPext=6.6756. High statistical significance of the 3D QSAR models has been achieved through the application of several data noise reduction techniques (i.e. smart region definition SRD, fractional factor design FFD, uninformative/iterative variable elimination UVE/IVE) on the original MIFs. Besides the ligand-based CoMFA models, an alignment molecular set constructed by flexible molecular docking has been also studied. The contour maps as well as the predicted reactivation potencies resulting from 3D QSAR analyses help better understand which structural features are associated with increased reactivation potency of studied compounds.