Flexibility of compounds in solution and our ability to represent it with conformational ensembles is of paramount importance in drug design. A compound’s behavior in solution is related to several important properties such as permeability (adaptability to polar and apolar environments) and the free energy cost of selecting a bioactive conformation. In an attempt to represent conformational flexibility for design purposes, computational chemistry studies use conformer generators ranging from knowledge-based approaches to quantum chemistry. A multitude of benchmarking studies of conformer generators against crystallographic databases has been published. Drawbacks of such benchmarks include the fact that they compare against conformers favored in a crystal lattice, rather than solution ensembles, and the fact that they do not recognize low abundance conformers, but rather label as “wrong” anything that doesn’t match the crystal structure.<br/>In this work, we look to evaluate several state-of-the-art conformer generators against solution data generated with ROESY NMR. Compounds for this study were selected for their flexible cores which were considered challenging. The NMR constraints extracted from this methodology are average distances between specific protons over the population of conformers in solution. Hence, the average distance over the computed conformers should be close to the NMR constraints if the ensemble properly represents the compound’s behavior in solution. The different conformer generators have been evaluated for how closely they reproduce the constraints on average, how many constraints are violated by more than a given threshold, how diverse the conformers are, and the cost-effectiveness of the algorithm.

Conformational flexibility in solution: Benchmarking conformer generators against a new ROESY NMR dataset

255th National Meeting (2018)

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