Identifying protein-protein interaction sites using granularity computing of quotient space theory

Abstract

The function of protein-protein interaction is very important to cell activity. Studying protein-protein interaction can help us understand life activities and pharmaceutical design. In this study, a kernel covering algorithm combined with the theory of granular computing of quotient space for predicting protein-protein interaction sites is proposed, (i.e. KCA-GS Model). This method achieves good performances, and the Sensitivity, Specificity, Accuracy and Correlation coefficient are 52.97%, 53.92%, 70.27%, 24.61%, respectively. It is indicated that our method is effective, potential and promising to identify protein-protein interaction sites. © 2010 Springer-Verlag Berlin Heidelberg.