Inferring protein-protein interactions from sequence using sequence order information

Identification of protein-protein interaction is crucial for nearly all biological process. In this paper, we introduce a new transformation of protein sequence based on sequence order information. We use 5594 interacting pairs of yeast organism from DIP core database as training data set, seven types organism as independent data set. The model with a new protein coding scheme obtains 88.927% accuracy, 88.243% sensitivity, 89.468% specificity, 77.864% MCC. The average performance on independent data set is 79.4999%.