A new method of screening molecules for their ability to bind to enzymes and other proteins could help discover potential new drugs.

A newly built model looks at covalent bonding, a much stronger type of chemical attraction but one that is harder to model. Computer models that predict which molecules will interact with enzymes have existed for a while, but they have been based on the shape of the molecules and weaker types of chemical attraction.

The new system has already found some candidate molecules that covalently bond to certain proteins involved in antibiotic resistance and in inflammation. Because of its potential usefulness, the team has made the system freely available on the web.

Original research paper published in Nature Chemical Biology on October 26, 2014.

Names and affiliations of selected authors

Brian Shoichet, University of Toronto, Ontario and University of California San Francisco, U.S.A.