Contest For Chemists

Three University of Cambridge chemists are challenging their fellow scientists to develop a model for calculating the solubility of a wide range of compounds.

"If you mention solubility, people won't immediately say, 'This is the forefront of science, and everyone should be interested in it,' " admits Jonathan M. Goodman, who, along with Antonio Llinàs and Robert C. Glen, published the challenge in the Journal of Chemical Information & Modeling (JCIM) (DOI: 10.1021/ci800058v). "But, in fact, they should be thinking that" because solubility can make or break the performance of newly designed drugs, polymers, foods, and other products, Goodman says.

Measuring solubility seems pretty straightforward, "but it's actually terribly difficult to do right," Goodman says. Time, temperature, pH, and other variables all affect solubility. As a result, solubility data in the literature aren't particularly reliable. For instance, the reported solubility of caffeine has increased by two orders of magnitude over the past century, he says.

Using a special titration technique developed in collaboration with Sirius Analytical Instruments of Forest Row, England, the Cambridge researchers have now determined the water solubilities of 100 druglike molecules, which they published in their JCIM article. The paper also lists 32 other molecules for which the researchers have measured, but not revealed, the solubility.

Scientists who enter the competition can use the database of 100 compounds to develop a solubility model and test how well it works by using it to predict the solubilities of the other 32 compounds. A successful model could then be used to predict solubilities for many additional compounds.

JCIM Editor William L. Jorgensen notes this is the first competition for the journal. He credits Glen, who is on JCIM's advisory board, and University of New Mexico pharmacist Anton J. Hopfinger, the journal's associate editor, with bringing the competition to fruition.

The journal editors will assemble a team to evaluate entries, which are due on Sept. 15. The contestants who score best will describe their modeling techniques in future issues of the journal. What's in it for the competitors? "Much as I would like to say huge amounts of money, what we're looking at here is scientific glory," according to Goodman.