A neural network has been used to predict both the location and the type of b-turns in a set of 300 nonhomologous protein domains. A substantial improvement in prediction accuracy compared with previous methods has been achieved by incorporating secondary structure information in the input data. The total percentage of residues correctly classified as b-turn or not-b-turn is around 75% with predicted secondary structure information. More significantly, the method gives a Matthews correlation coefficient~MCC! of around 0.35, compared with a typical MCC of around 0.20 using other b-turn prediction methods. Our method also distinguishes the two most numerous and well-defined types of b-turn, types I and II, with a significant level of accuracy~MCCs 0.22 and 0.26, respectively!.

No abstract

No abstract

Highlights► All the structural B-cell epitopes we examined are discontinuous. ► Only 18% of structural epitopes are spanned by a peptide fragment of 40 residues. ► Centralized and random distributions were considered for key functional residues. ► Fragments with only 7 residues will successfully span most key functional residues.