Intelligent computational method for discrimination of anticancer peptides by incorporating sequential and evolutionary profiles information

Cancer is one of the prominent threats to human life worldwide. Traditional therapeutic mechanisms like chemotherapy, radiation and surgical operations are exploited for its treatment. However, these clinical treatments are unfavorable, challenging and have severe impacts on human body. Recently, the discovery of anticancer peptides (ACPs) has become an influential anticancer drug agent due to their nontoxic characteristic and safe cellular uptake of therapeutic drugs. In this regard, much progress has been made to develop computational methods for ACPs prediction to accelerate their effectiveness against cancer. However, challenges remain in terms of discriminative feature representation, typical imbalance issue and prediction performance. In this study, we report a novel predictor, TargetACP, by integrating sequential and evolutionary-profiles information solely from primary protein sequences. Synthetic minority oversampling technique is utilized to cope with imbalance phenomenon between minority (ACPs) and majority (non-ACPs) samples. Finally, Support vector machine is employed as a learning hypothesis. Experimental results demonstrated that our predictor achieved an accuracy of 98.78% on benchmark dataset using jackknife cross-validation test. The generalization capability of the proposed method was evaluated through independent dataset which yielded accuracy of 94.66%. The empirical outcomes revealed that our model outperformed existing methods on same datasets. Furthermore, it is anticipated that TargetACP model will provide deep insights to pharmaceutical industry to design new anticancer drugs and research community to innovate new ideas in the area of bioinformatics, proteomics and computational biology.