- Article -
| In Silico Biology 3, 0035 (2003); ©2003, Bioinformation Systems e.V. |
Prediction of polypeptide fragments exposed to the solvent
Oliviero Carugo1,2
1 Department of General Chemistry of the Pavia University, Viale Taramelli 12,
I-27100 Pavia, Italy
2 International Centre for Genetic Engineering and Biotechnology, Padriciano 99, 34012 Trieste, Italy
Email: carugo@embl-heidelberg.de
Edited by E. Wingender; received May 13, 2003; revised and accepted June 07, 2003; published June 24, 2003
Abstract
A method is presented to predict those polypeptide segments within a globular protein that are more likely to be exposed to the solvent. The protein amino acidic sequence is the only information needed by this new algorithm. It uses a consensus hydrophobicity scale, derived from 28 known scales, and it is based on the comparison between the average hydrophobicity of a polypeptide fragment and the average hydrophobicity expected for a segment containing the same number of residues. The latter values are pre-computed from a non-redundant set of single chain protein structural domains. The comparison between the two average values results in a t value that readily provides the prediction with a statistical significance. A jack-knife validation analysis indicates that the protein segment predicted to be the most solvent exposed is actually solvent exposed and amongst the fragments that are most exposed. The source of a stand-alone program, written in C language, that allows the prediction of the most likely solvent exposed segment in a globular protein is available from the author.
Key words: accessibility prediction, epitope, hydrophobicity, modelling, protein structure, solvent accessibility, structure prediction