Prediction of Protein-Protein Interactions Using Symmetrical Encoding Scheme

The computational prediction of protein-protein interactions is currently an important issue in biology. In this paper, a K-local hyperplane distance nearest neighbor (HKNN) classifiers with symmetrical encoding scheme is proposed to predict protein-protein interactions. Moreover, a new sample encoding scheme, named symmetrical encoding scheme (SYES), for protein pair is developed by which a single protein-protein pair is mapped to two symmetrical points in the sample space. To evaluate the prediction performance of this encoding scheme, the ten-fold cross validation has been employed on two real data sets. The results indicate that this encoding scheme outperforms the sum encoding scheme to some extent, and the method proposed is comparable to other methods.