Consumable computing

Helping wet-lab chemists access the power of computational methods

Our research

The use of high-performance computing and simulation to support experimental science can greatly increase laboratory efficiency and provide additional insights into interactions not easily described by traditional methods. However, obtaining meaningful insight from simulations often involves significant investments in computational and human resources. This presents a major barrier to the widespread adoption of computational methods as a driver for laboratory exploration.

To overcome this consumability problem, non-experts should be able to obtain the results of cutting-edge computational science models as easily as they currently use a piece of wet-lab equipment. Such “computational appliances” would encapsulate the state-of-the-art computational know-how, automate the generation of final results from a few input parameters, and be accessible via a user-friendly interface. The user’s interaction with these appliances would focus on the scientific functions they provide rather than the underlying technology they run on.

At the Hartree Centre, IBM, STFC with our industrial partners, specially their wet-lab scientists, we have developed in silico counterparts to laboratory experiments the wet-lab scientists commonly perform as part of their R&D activities. Currently we have developed a number of proof-of-concept appliances accessible via an iPad app. The goal of this project is to demonstrate this technology in operational industrial environments by providing experimentalists with access to production computational appliances, which they will use to augment their work.

Watch the phase separation demo

This appliance examines mixtures made of different proportions of a selected molecule (xylene or toluene) with methanol and water and determines if the molecule dissolves in the mixture or separates out. The result is shown on a ternary-phase diagram (green points: dissolved, red points: separated) with the estimated boundary line between the different behaviour regions (the phase boundary) shown. The application uses a cognitive algorithm to intelligently sample the phase diagram, using the values of completed points to decide where it would be most efficient to sample next.

Learn more about predicting ternary-phase diagrams from molecular simulation.

Ask the experts

Michael Johnston

Michael Johnston
IBM Research – Dublin, Ireland

James Mcdonagh

James Mcdonagh

Kirk E. Jordan

Kirk E. Jordan
IBM T.J. Watson Research Center