The marriage of state of the art GPU computing resources and physics-based algorithms to accelerate drug discovery projects

The holy grail of computer-aided drug design is the accurate prediction of the binding affinity of a drug molecule for its protein target. If this were possible, even the prediction of relative, rather than absolute, binding affinities, the impact on drug design would be phenomenal as it would enable scientists to zero in rapidly on the key compounds to make, and to discard poor candidates without having to synthesize them. In recent years, some progress towards this goal has been made as massive computing resources based on high-end graphics cards (usually used in computer gaming) have been applied to simulations of protein-ligand complexes allowing better estimates of binding affinities … but only in some cases. Anecdotally, the technology has been shown to work in some cases, but not all – and it’s not possible to tell beforehand which is which. But it’s still relatively early days and the hope is that with growing experience and continuing development of the underlying methodologies, the success rate of this technique will improve.

Our What’s Hot forecasts will appear on Eureka through December. Follow along and feel free to share your own predictions for the new year.