Reducing internal strain energy in small molecules is critical for designing potent drugs. Quantum mechanical (QM) and molecular mechanical (MM) methods are often used to estimate these energies. While in general, QM methods are slow and more accurate and MM methods are fast and less accurate, which methods offer an optimal balance? To aid in answering this question, we have chosen a test set of 62 drug-like fragments and carried out torsion scan analysis. We compared potential energies calculated using 6 QM and 6 MM methods to QM calculations at a higher level of theory, Moller-Plesset, MP2. Comparing QM methods with the same basis set (6-311+g**), we found Hartree-Fock (HF) exhibits a smaller deviation in energy from MP2 (dE=0.5 kcal/mol mean) compared to the widely used B3LYP (dE=0.7 kcal/mol) which in turn is better than three semi-empirical methods, AM1, PM3, and PM6 (all with dE=1.3 kcal/mol). Among MM methods, the OPLS2.1 forcefield is more accurate (dE=0.7kcal/mol) compared to the commonly used MMFF94S forcefield (dE=1.8 kcal/mol). Following the HF results, we present a new torsion scan protocol whereby a HF geometry optimization with a smaller basis set is followed by single point energy calculation using MP2. We also point out the limitations in a number of methods with the aim of providing motivation to improve forcefields.

Comparison of industry-standard QM and MM methods for estimating strain in drug-like molecules

251st National Meeting (2016)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.