HWI is dedicated to improving the health and well-being of mankind through excellence in basic biomedical research. The institute focuses on improving drug design through crystallographic research. In recent years, a team of HWI scientists, including C.M. Weeks, R. Miller, and Nobel prize-winning chemist H.A. Hauptman have developed a computer programme called Shake-and-Bake (SnB), which uses a state-of-the-art algorithm for solving crystal structures by ab initio, dual-space methods. SnB helps researchers determine the 3D structures of particularly difficult macromolecules such as proteins, from crystallographic data.
X-ray crystallography helps the scientists to understand not only the 3D structure of target macromolecules, but also the way in which a specific drug may interact with selected macromolecules in the body. The researcher bombards a crystalline sample of material with X-rays, watching to see how the atoms scatter the X-rays. Using this process, the researcher can discover how a potential drug interacts with its targets, enabling him to improve lead compounds, reduce side-effects, and increase the effectiveness of new drug therapies.
Using molecular information from X-ray crystallography is sometimes the only clue which scientists have of what to do next in the drug development process and this type of work wouldn't be possible without the continuous development of computers, according to Miller. Up until this decade, people were just working with diffraction data and trying to determine the actual value of the phases. Today, scientists can analyse how different molecules interact with each other, as Weeks states. With this knowledge, they can develop a better drug, perhaps tuning it so that it interacts with a particular target even better. However, this type of work is much more computationally intensive, according to Weeks.HWI gives universities and for-profit organizations access to SnB so they can perform x-ray crystallography for their own research and drug development. The motivation in setting up the AlphaServer is to solve the issue of people wanting to use the SnB programme but without having the computational resources to run it, as explains Chuck Weeks, who is co-Director of the SnB research team. Some of these companies have very proprietary information to protect, so security is a major concern, he adds.
To accomplish this goal, HWI needed a robust, reliable, as well as powerful system to run SnB, a firewall, and a way to provide secure communications over the Internet, such as tunnelling or VPN. Russ Miller, SnB co-Director of the project, decided to select Compaq because of the promise of being able to work with Alta Vista security. He wanted a turnkey solution to be able to put the code in and get off and running. No other vendor was able to offer this. In collaboration with Compaq account manager Scott Fisher, sales support consultant Tom Knight, and Compaq distributor Dartnell, HWI ran a full line of benchmark tests at the Toronto Benchmarking Center.
Miller's team has evaluated a variety of platforms and operating systems and compilers. The Alpha system had the fastest processors and most stringent compilers. The team was very satisfied with the Alpha performance, quality of compilers, and the ease with which it could move the programmes from other platforms to Alpha. HWI purchased a four-processor AlphaServer 4100 5/466, 1 GB of memory, six 4.3 GB disk drives, plus an AltaVista Firewall and Tunnel for secure communications on a separate DIGITAL Personal Workstation, all running the Tru64 UNIX operating system.
HWI now has a very secure system behind their firewall, as well as secure communications via tunnelling. Outside users gain access to SnB via a link on HWI's Web site, and those who have security needs can access SnB via AltaVista tunnelling software. HWI debuted the availability of SnB at the annual American Crystallographic Association Meeting in May 1999. They're currently using the AlphaServer internally for further application design. The team hopes the demand for the machine will be so successful that it will no longer be able to use the machine for its own purposes.
As more data becomes available from genome projects, HWI scientists will have access to databanks of genetic information. As such, they will be able to understand what particular genes are coded for and what specific proteins do. HWI will need to determine the structures of these proteins as a follow-up to the genome project and hopes to play a major role in that effort.