Data base requirements for graphical applications in biochemistry

Abstract

Although there are numerous areas of the life sciences that could benefit from having a general and universal data base for three-dimensional computer display systems, perhaps the most apparent is for structural studies of biological macromolecules (molecular graphics). Since the structure of the first protein (myoglobin, which contains 1400 non-hydrogen atoms) was solved by Sir John Kendrew and coworkers in 1962 by x-ray diffraction methods, no fewer than 80 different molecules of that size or larger have been solved. Probably in the next 5 years another 200 to 400 additional structures will be solved. These will include not only proteins, nucleic acids and polysaccharides, but complexes of these as well, where the total molecular weight may reach 10 to 50 times that of myoglobin (e.g. complete virus particles). The ultimate goals of the investigators are to discover the mechanism of action of these molecules and complexes, to predict three-dimensional structure and function of hypothetical molecules, and finally to be able to synthesize new macromolecules of new and predictable functions. In the past several years methods for solving large structures and refinement of models to observed crystallographic data have improved dramatically and will notably improve the accuracy of such models. It is quite clear that this improved accuracy will greatly aid reaching these objectives. Furthermore, these goals will only be possible if the present and future structural information can be thoroughly studied and assimilated. Computer graphics provides a reasonable hope and a sophisticated data base is an obvious necessity. Data base requirements for molecular graphics, not only those for display, manipulation and solving complex structures for they have been well demonstrated, but also those performing research studies on the accumulated results, are certainly within the grasp of current methods.