ABSTRACT

Crystallography has already revealed a great number of biopolymer structures at full atomic resolution, and the productivity of structural biologists is currently increasing at a breathtaking pace. The enormous amounts of data collected in structural databanks contain a true wealth of information. They are readily used in discussions of catalytic mechanisms of enzymes and ribozymes and provide the basis for models of molecular recognition. Many other applications of structural data in biochemistry and molecular biology, however, require fewer details and thus call for notions of coarse-grained structure. Too many data obscure common structural features in related biopolymers and impede comparisons that are of fundamental importance, for example, in molecular evolution. Discretized structure models are particularly interesting because they not only meet the need for straightforward recognition of basic features but also by their nature they can be enumerated and accessed by combinatorial and other rigorous mathematical techniques.