Title: Accuracy of protein hydropathy predictions

Authors: Satu Jaaskelainen, Pentti Riikonen, Tapio Salakoski, Mauno Vihinen

Addresses: Department of Information Technology, University of Turku, FI-20014 Turku, Finland; Bioinformatics Laboratory, Turku Centre for Computer Science, FI-20520 Turku, Finland. ' Department of Information Technology, University of Turku, FI-20014 Turku, Finland; Bioinformatics Laboratory, Turku Centre for Computer Science, FI-20520 Turku, Finland. ' Department of Information Technology, University of Turku, FI-20014 Turku, Finland; Bioinformatics Laboratory, Turku Centre for Computer Science, FI-20520 Turku, Finland. ' Institute of Medical Technology, FI-33014 University of Tampere, Finland; and Tampere University Hospital, FI-33520 Tampere, Finland

Abstract: Hydropathy is a dominant force in protein folding. Sequence-based hydropathy predictions are widely used, without knowledge about their accuracy and reliability. We investigated the prediction accuracy of 56 hydropathy scales by correlating predicted values with the accessible surface area in known protein structures. Results for different amino acids vary greatly within each scale. We also investigated prediction accuracies of amino acids separately in secondary structural elements and in protein fold families. Despite very low overall correlation, hydropathy predictions can still be used if the shape of the plot is important instead of the prediction values.

Keywords: prediction accuracy; amino acid solvent accessibility; accuracy assessment; protein hydropathy predictions; bioinformatics; protein folding; amino acids; protein structures.

DOI: 10.1504/IJDMB.2010.037550

International Journal of Data Mining and Bioinformatics, 2010 Vol.4 No.6, pp.735 - 754

Received: 20 May 2009
Accepted: 04 Sep 2009
Published online: 16 Dec 2010 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article