Prediction of anticancer peptides with dictionary learning method

Recent years anticancer peptides, with advantages of non-side effects and well therapeutic effects, has become a potential treatment for cancer worldwide. However, acquisition of anticancer peptides in clinical practice is time and resource consuming due to low successful translational rate. In this study, we propose a novel method of predicting anticancer peptides based on dictionary learning (DLACP) which can detect the target peptides in a short time with high-precision. Specifically, in order to obtain the effective representative information, the proposed method utilize dictionary learning to encode physicochemical properties of the proteins by clustering them from amino acid indices (AAindex) database. Experimental results show that the proposed method can achieve better performance of predicting anticancer peptides that cutting-edge methods with a considerable margin gap.