G-RMSD: Root Mean Square Deviation Based Method for Three-Dimensional Molecular Similarity Determination

We present the Generalized Root Mean Square Deviation (G-RMSD) method. G-RMSD is an optimization method to calculate minimal RMSD value of two atomic structures by optimal superimposition. not restricted systems with equal number atoms compare or a unique atom mapping between molecules. The can handle any type chemical structure, including transition states and which cannot be explained only valence bond (VB) theory (non-VB structures). It requires Cartesian coordinates for structures. Further information, i.e. atom- types also included. Applications classification α-d-glucose conformers 3D partial structure search using dataset containing equilibrium (EQ), dissociation channel (DC), state (TS) are demonstrated. find that allows successful wide variety molecular G-RMSD, (Root Deviation) superimposition, solves problem comparing 3D-molecules different size without mapping. was successfully applied searching quantum mechanical based reaction pathway database structures, but non-valence