Computational prediction of the specificities of proteasome interaction with antigen protein

In the processing and presentation of antigenic peptides bound by the major histocompatibility complex (MHC) class I molecule, the ubiquitin-proteasome system of the eukaryotic cells plays an important role in proteolysis and degradation. The ubiquitinated protein substrate is delivered into the 26S proteasome to be digested and degraded. The proteasome degrading substrate is actually protein-protein interactions. Some researches of predicting proteasome cleave site rarely gave the information of the proteasome interacting with its substrate, and so the accuracy and reliability of these proteasome cleavage predictive methods still need to be improved. This paper used support vector machine method (SVM) to predict the proteasomal cleavage sites, and the predictive accuracy of the model was 82.8%. We showed analytically that the cleavage specificities of the cleavage sites "|" and its adjacent positions, and gave the information about the proteasome interacting with its substrate from our research results. It demonstrates that the proteasome cleaving to target protein is selective, but not random.