Protein Structure Classification

Abstract

To understand and map the universe of protein structures, it is necessary to collate, annotate and classify these structures in a rational scheme. While the individual protein structures reveal much about the specific molecular mechanisms that underlie a particular biological function, taken together this body of data also allows biologists to explore the evolution of structure and function. As structure is much better conserved than sequence, these data also facilitate the recognition of evolutionary relationships that are hidden at the sequence level. The different approaches that have been taken to tackle this problem include the identification of protein domains, phylogenetic and phenetic classification and hierarchical and nearest-neighbour clustering. Powerful sequence searching methods then enable structural assignments to be allocated to genomic data. Key Concepts Proteins comprise recognisable smaller sequence domains. Domains usually consist of secondary and supersecondary structures and have an average size of 150 ± 50 residues. These domains may be thought of as units of evolution, which recur in many proteins in various combinations. Structural classifications have been developed that group domains into fold and sequence families. The phenetic approach groups the proteins according to their structural characteristics. The phylogenetic approach groups proteins into families according to their evolutionary history. Keywords: protein structure classification; common folds; protein architecture; structural comparison; genomes