A Simple Comparison between Specific Protein Secondary Structure Prediction Tools

A comparative evaluation of five widely used protein secondary structure prediction programs available in World Wide Web was carried out. Secondary structure data of ten proteins containing 190 secondary structure motifs were collected from Protein Data Bank (PDB). The amino acid sequences of the proteins were then evaluated using GOR, PSIPRED, HNN, PROF, and YASPIN secondary structure prediction tools and the results were compared with the structural information obtained from PDB. The study reveals considerable differences between results obtained from each program. Within the limit of this comparative study, PSIPRED showed the highest prediction accuracy with 77 % accuracy in α helix prediction and 70 % accuracy in b strand prediction. Furthermore, the level of accuracy varied with the length of the secondary structure motifs. Highest accuracies were obtained for α helices of 16-20 amino acids and β strands of 7-9 amino acids in length. The results suggest that, among the most frequently used software programs available in World Wide Web, PSIPRED is the tool that gives the best results for secondary structure prediction. Tropical Agricultural Research Vol. 23 (1): 91-98 (2011) DOI: http://dx.doi.org/10.4038/tar.v23i1.4636