Published in

Royal Society of Chemistry, Chemical Science, 5(2), p. 883

DOI: 10.1039/c0sc00555j

Links

Tools

Export citation

Search in Google Scholar

Virtual cocrystal screening

This paper is made freely available by the publisher.
This paper is made freely available by the publisher.

Full text: Download

Red circle
Preprint: archiving forbidden
Green circle
Postprint: archiving allowed
Green circle
Published version: archiving allowed
Data provided by SHERPA/RoMEO

Abstract

Calculated gas phase molecular electrostatic potential surfaces have been used to identify sets of H-bond donor and H-bond acceptor sites that describe the possible intermolecular interaction sites on the surface of a molecule. The calculated H-bond parameters, αi and βj, were used to estimate interaction site pairing energies in the solid form of the compound through a hierarchical mapping of complementary donor and acceptor sites: the interaction energy for each contact is simply given by the product −αiβj. The approach assumes that all of the interactions that can be made in the solid are made and that the details of three-dimensional structure and crystal packing are of secondary importance. Comparison of the energy of two pure solids with cocrystals of various stoichiometries gives an energy difference, ΔE, which is a measure of the probability of forming a cocrystal. Tests on an experimental cocrystal screen from the literature and the recall of coformers for caffeine and for carbamazepine from a list of nearly 1,000 candidates have been used to validate the utility of the method. For systems that are experimentally found to form cocrystals, the calculated energy parameter ΔE tends to be very favourable. In the best case, for 846 potential caffeine coformers from the EAFUS list, 80% of the experimentally observed hits are in the top 11% of the ranked list of ΔE values. The results provide a calibration between the value of ΔE and the probability of cocrystal formation: when the cocrystal is favored by more than 11 kJ mol−1 over the two pure solids, the probability of obtaining a cocrystal is better than 50%. An advantage of this approach is that it is sufficiently fast to be used as a high throughput virtual screening tool.