Protein function and inhibitor prediction by statistical learning approach

Abstract

Protein function prediction system consists of a number of protein functional families such as enzyme families, transporter families and RNA-binding proteins were studied and a protein function classification system was designed for assigning functional families irrespective sequence similarity. In addition, novel proteins were selected and tested to examine to which extent, their function can be predicted by using our prediction systems. The evaluation results strongly suggest that an SVM-based prediction system is useful for facilitating the prediction of the function of distantly related proteins in the genomes of bacteria, virus, plants as well as other organisms. Another aim of my work is to predict protein inhibitors by statistical learning approach and prediction of HIV-protease inhibitors (PIs) was used as an example. The results indicated that the statistical learning approach is useful for PIs prediction, the approaches used here could be extended to the other inhibitor/agonist/substrate prediction problems.