Predicting Protein-Protein Interactions Using Correlation Coefficient and Principle Component Analysis

A new features for predicting protein-protein interaction with neural classification is proposed. Our feature extraction is based on the correlation coefficients of physicochemical properties and the statistical means and standard deviations of five secondary structures, i.e. alpha-helix, beta-sheet, beta-turn, coil, and parallel beta strand. The proposed method is tested with yeast Saccharomyces Cerevisiae proteins. Our result uses fewer features which is 50% less than the other´s and achieves 92.15% accuracy higher than the other other´s.