'Cyber-enabled' Chemistry

Advances in technology can be powerful drivers. Take, for instance, the progress in computer, information, and communication technologies, which have merged to produce a new, comprehensive cyber-infrastructure. It's this infrastructure that is generating excitement at the National Science Foundation: The infrastructure can provide a platform upon which scientists and engineers can build to create integrated databases and remote experimental control environments with the potential to more efficiently tackle difficult scientific problems.

At the forefront of this focus area is NSF's Chemistry Division. The division has been working to promote cyber-enabled chemistry-the application of this cyber-infrastructure to enable new chemical research and education activities through grid computing, community databases, remote access to instrumentation, electronic support for geographically dispersed collaborators, and other Web- and grid-accessible services.

Cyber-enabled chemistry will provide new paradigms for chemical research and education, NSF Chemistry Division Director Arthur B. Ellis explains. Imagine operating instruments remotely from your desktop, accessing a complete suite of chemical databases and data mining tools, and utilizing molecular visualization and grid-computing capabilities so that you can seamlessly combine experiment, theory, modeling, and simulation to tackle research problems previously thought to be intractable. Or imagine data-driven and simulation-informed experiments that enable the computer to direct the instrument or synthetic apparatus to adjust an experiment in real time and to guide the choice of the next experiment, he says.

To help encourage work in cyber-enabled chemistry, the Chemistry Division, in cooperation with what was formerly the NSF Division of Shared Cyberinfrastructure, developed the Chemistry Research Instrumentation & Facilities: Cyberinfrastructure & Research Facilities (CRIF:CRF) program. The $10 million in awards is designed to lay a solid infrastructural base for future chemical research and education building on the vision identified by the chemical community at last fall's cyber-enabled chemistry workshop (C&EN, March 14, page 3).

The use of information, computational, and communication technologies is now an accepted part of the chemical research infrastructure, points out Celeste M. Rohlfing, program director and cyber-enabled chemistry contact person in the Chemistry Division at NSF. Yet concerted efforts are required to improve software interoperability, to prototype modeling to a prescribed accuracy with well-defined benchmark experiments and simulation data, to archive and warehouse experimental and simulation data, and to improve collaboration tools. This program targets those efforts, she notes.

To that end, the Chemistry Division has announced the first set of grants under the CRIF:CRF program. According to Rohlfing, the division received 18 proposals, all of which were multi-investigator and most of which involved multiple institutions.

Although the proposed projects are expected to serve state-of-the-art software needs for a wide community of users in chemistry, several of the submissions were quite diverse with respect to other disciplines encompassed-specifically, at the interfaces with biology, materials, engineering, and geosciences, Rohlfing points out. This diversity does not preclude them from being funded through the program so long as they substantially advance the chemical sciences, as determined by merit review, and are centered in traditional chemistry subdisciplines or serve as interfaces with other fields, she says.

In September, the division selected four projects to fund. The principal investigators (PIs) from these selected projects are Teresa Head-Gordon, associate professor of bioengineering at the University of California, Berkeley; Todd J. Martinez, professor of chemistry at the University of Illinois, Urbana-Champaign; Karl Mueller, associate professor of chemistry at Pennsylvania State University; and Michael Frenklach, professor of mechanical engineering at UC Berkeley.

For their part, the awardees are excited about the potential of cyber-enabled chemistry. Chemistry consists of a huge number of experiments and calculations, each done on a limited number of specific molecules in particular conditions, says William H. Green Jr., associate professor of chemistry at Massachusetts Institute of Technology and co-PI with Frenklach on this grant. Frequently, the measurements or calculations are performed in laboratories scattered over different countries, different decades, and use different protocols based on different assumptions, he notes.

Progress in chemistry occurs when someone correctly generalizes from this set of disparate data to come up with a general principle that correctly predicts how these and other related molecules would behave under some reaction conditions where they have not been measured, Green explains. Cyber-tools, by enabling scientific methods, he says, will provide means to develop scientific generalizations tested against all data that are available.

Pulling all of these data together in a meaningful way is a challenge faced in any field, but particularly in Frenklach's area of study-gas-phase chemical kinetics. To tackle this barrier, he is leading a multiple-institution group of collaborators working on a project known as the Process Informatics Model (PrIMe) initiative. According to Frenklach, this initiative-which has evolved from a project he was part of more than 10 years ago-aims to create cyber-infrastructure and cyber-tools for the development of predictive reaction models from a collection of available data.

The team led by Penn State's Mueller is also working to bring together large bodies of data. Within our project, Mueller says, cyber-infrastructure tools will be designed to enable and promote multidisciplinary and multiscale research in environmental kinetics analysis and synthesis.

The CRIF:CRF program fosters this unique environment for collaboration between chemical researchers and information scientists, Mueller explains. Bringing these two groups of researchers together with different approaches but a common goal engenders a unique attack of the data overload' problem, he notes. Then, he adds, we can look for new patterns in the data that may only emerge when a large amount of information is viewed in a unique way.

The collaborative nature of the program goes beyond just chemistry and information scientists. Researchers are well-equipped for communication within their disciplines, but making connections-real and virtual-among chemists, geoscientists, environmental engineers, and other scientists is critical for breaking down disciplinary barriers to collaboration in environmental kinetics analysis, Mueller says, adding that his team is seeking to improve such multidisciplinary communication.

Martinez and his colleagues are also pulling together multiple disciplines to tackle chemical problems. The primary goal of his team's work is to make first-principle quantum dynamics methods widely available and easy to use.

Cyber-enabled chemistry consists of new approaches that exploit the presence of ubiquitous, practically unlimited network, information, and computing resources to solve chemical problems, Martinez says. The effective utilization of cyber-resources will allow us to take a more problem centered' or systems-level' approach to complex questions that have so far resisted solution by any single method or approach, he notes.

Another area of chemistry that will benefit from this set of grants is biomolecular modeling. Head-Gordon's team is working to improve theoretical model infrastructure for biomolecular simulations. The project includes development and validation of polarized force fields for use in these simulations and the sharing of resulting software components and parallelized computer codes with the community.

In addition to engaging chemical researchers, the NSF program also offers outreach opportunities. Our community has outstanding opportunities for engaging the public in citizen chemistry,' by encouraging participation, for example, in distributed computational experiments whose outcomes would be of broad general interest because they relate to themes like energy, health, the environment, or the economy, Ellis notes.

Ellis also points out that while the CRIF:CRF program funds relatively large projects in cyber-enabled chemistry with teams of researchers, there are opportunities within other NSF grants to contribute to this area. We also want to encourage individual PIs to cyber-size' their projects, by including funds as appropriate in their budgets to support the use or development of cyber-tools that can be made broadly available to the chemical sciences community, he explains.

As the Chemistry Division prepares for the next round of these grants (the solicitation is expected to be posted later this year), Ellis and Rohlfing look forward to seeing this area address broad chemical challenges in research and education. Cyber-enabled chemistry is only limited by our imagination, Ellis says.