Automatic recognition of domains in globular proteins

Abstract

Publisher Summary This chapter describes a method to reveal the folding pattern, but the analysis stops short of introducing additional physical criteria that would be needed to identify particular units within the pattern. Nevertheless, initial steps in the current analysis often result in chain divisions that correspond to the typical large domains as seen by crystallographers. The algorithm described is devised to optimize sequential subdivisions of the folded polypeptide chain into spatially distinct regions without regard for chemical particulars such as residue type, hydrogen bonding, or disulfide loops. This chapter relates the crystallographer's vantage plane (disclosing plane) to an intramolecular partitioning plane, and describes an algorithm to find both. When this is done, it becomes apparent that domains as such are not limited to the topmost echelons of molecular structure; instead, they are reiterated in hierarchic fashion ranging from the whole protein monomer through supersecondary structures down to individual helices and strands. The concept of a domain is subject to reinterpretation when considered in the broader context of its molecular hierarchy. The conventional large domains are seen to be structural composites, with subparts that are domains in their own fight.