Boston University Chemists, Computer Scientist Win NSF Grant For Computational Chemistry Project

Contact: Ann Marie Menting, 617/353-2240 |

(Boston, Mass.) — A three-year, nearly half-million dollar grant from the Information Technology Research (ITR) program of the National Science Foundation (NSF) will spur efforts by three Boston University scientists to develop time-saving computational approaches to predicting how liquids form glasses and how proteins fold. The researchers also hope to spark interest among BU undergraduate and graduate students in the field of computational chemistry, a discipline rich in interesting chemistry, challenging computational problems, and promising careers. Student recruitment for this project is underway.

Grant recipients Thomas Keyes and John Straub, professors of chemistry and members of BU’s Center for Computational Science (CCS), and Steven Homer, professor of computer science and member of CCS, aim to build computational tools that can simulate on an accelerated timescale the step-by-step processes that occur in ultraslow systems such as biomolecules and supercooled liquids. These computational tools, known as algorithms, will vastly shorten the time needed to predict how proteins fold into the active forms needed in various biological processes or how liquids that are cooled quickly to temperatures below their normal freezing point form glasses that are used in myriad products.

“Simulating these processes using conventional methods requires more steps than the national debt in pennies,” says principal investigator Keyes, “which means that the most important problems cannot be studied by even the largest supercomputers. However, most of the steps do not contribute to the important changes, so simulating them all is a great waste of computational time.

“Our goal is to focus on points of large-scale molecular motion and to build algorithms that will just make those big moves, without waiting to calculate the in-between, small steps in the natural movement of the molecules.”

Simulating systems that contain both large-scale slow processes and small-scale fast processes means tackling what is known as the multi-timescale problem, an outstanding challenge for theoretical chemists such as Keyes and Straub. The project will address this challenge using what Keyes describes as potential-energy landscape approach to mapping molecular movement.

The energy of interaction of the molecules — their potential energy — depends on their position or arrangement. This can be imagined as a rough or mountainous landscape in which molecules go slowly through valleys but then quickly move over the mountains. Folding a protein or forming a glass involves crossing several mountain ranges.

Conventional simulations cause the systems to wander around the valleys, which eats up computation time. The algorithms being developed by the team of researchers will first map the landscape and then move the systems through a series of steps that only include mountain crossings, thus speeding the process and cutting the computational time.

NSF’s ITR program was instituted to foster research and education that advances in information technology. Grants made in fiscal year 2003 targeted a few key areas, most notably projects that would expand the use of information technology across the sciences and engineering and create novel uses for the technology.

Boston University’s Center for Computational Science is exceptionally equipped to address such goals. Established in 1990, the Center coordinates and promotes computationally based research, fosters computational science education, and provides computational resources and support to University researchers.

Researchers in the Boston University’s Department of Computer Science address questions of computation, programming language, and information theory as well as cryptography, algorithm analysis, networking and Internet systems, and image, video, and real-time computing. In addition to theoretical research, scientists in the Department of Chemistry investigate questions in chemical physics, photochemistry, inorganic and organic chemistry, physical chemistry, and biochemistry.

Boston University, the nation’s fourth largest independent university, has an enrollment of more than 29,000 in its 17 schools and colleges.