Statistical confidence in molecular docking: a study with GluN1/GluN2B NMDAR ligands

Redocking studies show that scoring data do not necessarily correlate to better RMSD of the related poses. Since screening studies are generally based on scoring data, how reliable are these results? In this work, we explore the molecular docking of ifenprodil and some analogues related to the treatment of bipolar disorder at GluN1/N2B NMDAR phenylethanolamine site1,2,4. The NMDAR is associated with glutamatergic system deregulation in bipolar individuals, most with depression mood1,2,3,4. Starting from NMDAR/ifenprodil co-crystal (PDB: 3QEL) 2, based on RMSD values from 3 scoring functions of the GOLD software, data generated from a redocking run (population size of 1000 solutions) were used to standardize the docked series5. We were able to define the most reliable score range that enabled the group of poses was most correlated to the experimental ligand poses. This can improve screening of databases in the search of new candidates for the treatment of bipolar disorder.