ITR: Enhancing Crystal Structure Determination through Data Mining, Collaborative Environments, and Grid Computing

ITR: Enhancing Crystal Structure Determination through Data Mining, Collaborative Environments, and Grid Computing X-ray crystallography, with its unique ability to reveal the atomic or near-atomic structures of a wide range of biomedically important molecules, is the cornerstone of modern structural biology. This project will support critical advances in the enabling technologies of automated tuning of software, learning via data mining, geographically distributed collaborative environments, and grid computing for the enhancement and determination of crystal structure. Computer programs based on "direct methods" are used to determine the majority of small-molecule organic crystal structures. These methods begin to fail in the 100-200-atom range because the accuracy of the underlying probabilistic relationships is inversely proportional to the square root of the size of the structure. The Shake-and-Bake algorithm and SnB program have extended the size of crystal structures amenable to direct-methods phasing from 100 to 2500 atoms. SnB has also been used to increase the size of heavy-atom substructures in large proteins that can be determined from 10 to 180 Se atoms. This provides a bootstrap by which complete structures, containing hundreds of thousands of atoms, can be elucidated. Such accomplishments would have been regarded as impossible only a few years ago, and the ultimate potential of the Shake-and-Bake approach to ab initio structure determination of macromolecules is unknown. The integrated SnB collaborative environment will allow multiple simultaneous users at distinct locations to work together in a virtual environment via a variety of platforms (from 3D fully immersive to desktop PC). Users will be able to navigate through a structure, search and import structures available from public crystallographic data banks, edit structures, and interface directly with SnB while running either in solution mode or in refinement mode. The users will have the ability to work in a collaborative fashion as they would if they were all situated in close physical proximity. Finally, platforms consisting of computational grids are available in many academic and commercial institutions that use crystallographic software. A grid-enabled version of SnB will be created so that it can take advantage of computational grids and networks of workstations that have available computing cycles. The introduction of SnB has had an enormous impact on the crystallographic community. The integration of the Shake-and-Bake methodology with automated data warehousing and data mining should provide equally spectacular advances in the near future. The introduction of collaborative environments and the ability to exploit computational grids is expected to have a significant impact as well. Advances in enabling technologies coupled to a widely-distributed package, provides a unique opportunity to evaluate the viability of these enabling technologies, which have far-reaching applications to a wealth of diverse areas, including computational chemistry, advanced design, optimization, and edutainment.