Automatic identification of secondary structure in globular proteins.

A computer program is used to analyse automatically and objectively the atomic co-ordinates of a large number of globular proteins in order to identify the regions of a-helix, p-sheet and reverse-turn secondary structure. Several different criteria for the assignment of secondary structure are tested for accuracy, reproducibility and efficiency. The most successful criterion, which is based on patterns of peptide hydrogen bonds, inter-0 distances and inter-@ torsion angles, is used to find the secondary structure of all the proteins studied. The accuracy of the derived assignments is assessed by comparing them with the secondary structure reported in the literature for each protein. The reliability of the methods is . the independently assessed by comparing the secondary structures derived from determined sets of co-ordinates available for some proteins. We provide the first objective and consistent compilation of a-helix, p-sheet and reverse-turn secondary structure in almost all globular proteins of known tertiary structure. These data will be invaluable for analysinlg the relative tendencies of different amino acids to occur in different types of secondary structure, for analysing the regularity how the pieces of secondary structure of each protein. of the secondary structure itself, and for analysing structure fit together to form the globular tertiary