DETERMINING MOLECULAR STRUCTURES
Cornell University Ithaca, Ithaca NY
Investigators
Linked publications & trials
Abstract
Rational drug design depends on having knowledge of the three-dimensional shapes of potential drugs and other molecules involved in disease processes. Unfortunately, no microscope is capable of providing direct images of biologically significant molecules. However, it is possible to "see" such molecules by using the techniques of X-ray crystallography. Traditional direct methods techniques permit crystallographers to mathematically reconstruct the three-dimensional structure of the molecules responsible for a given diffraction pattern provided that the molecules contain no more than approximately 100 atoms. Unfortunately, larger structures, including protein molecules that interact with drugs in living organisms, have traditionally been determined by more labor-intensive methods requiring chemical modification and the measurement of multiple diffraction patterns. In fact, no generally applicable method existed at all for molecules containing 100-500 atoms until a new procedure known as Shake-and-Bake (Miller, et al., 1993; Weeks, DeTitta, Hauptman, Thuman & Miller, 1994) was created. During this past year we have used the IBM SP in order to solve a (previously known) 1000+ atom structure, Lysozyme, with our SnB (Shake-and-Bake) program. Breaking the 1000-atom barrier represented a milestone in direct methods research. We have recently created a parallel version of a pre-release of SnB 2.0, the latest version of our program, which is currently being tested on the IBM SP. We expect to be using this machine to attempt to solve an unknown, 1600-atom structure, by the end of 1997. Given enough resources, we hope that a solution will be at hand within several months.
View original record on NIH RePORTER →