A Short Review on Protein Secondary Structure Prediction Methods

• Published in: Pattern Recognition in Computational Molecular Biology pp. 97 - 113
• Main Authors: Yan, Renxiang, Song, Jiangning, Cai, Weiwen, Zhang, Ziding
• Format: Book Chapter
• Language: English
• Published: Hoboken, NJ, USA John Wiley & Sons, Inc 19.11.2015
• Subjects: Chou‐Fasman method; Garnier, Osguthorpe and Robson method; pattern recognition algorithms; PHD method; protein secondary structure prediction; SPINE‐X
• ISBN: 9781118893685; 1118893689
• DOI: 10.1002/9781119078845.ch6

This chapter discusses seven protein secondary structure prediction methods, covering simple statistical‐and pattern recognition‐based techniques. The prediction methods include Chou‐Fasman, Garnier, Osguthorpe and Robson (GOR), PHD, neural network (NN)‐based protein secondary structure prediction (PSIPRED), SPINE‐X, protein secondary structure prediction (PSSpred) and meta methods. The chapter assesses the performance of different methods using the Q 3 measure. It investigates the accuracy of secondary structure prediction for target proteins by the alignment/threading programs. The top‐performing methods, for example, PSSpred, PSIPRED, and SPINE‐X, are consistently developed using NNs, which suggests that NNs are one of the most suitable pattern recognition algorithms to infer protein secondary structure from sequence profiles. It is found that the secondary structure prediction by the alignment/threading methods that combined PSIPRED with other informative structural features, such as solvent accessibility and dihedral torsion angles, was more accurate.