A Novel Hybrid Sequence-Based Model for Identifying Anticancer Peptides.

Lei Xu,Longjie Wang,Changrui Liao,Guangmin Liang

doi:10.3390/genes9030158

Abstract

Cancer is a serious health issue worldwide. Traditional treatment methods focus on killing cancer cells by using anticancer drugs or radiation therapy, but the cost of these methods is quite high, and in addition there are side effects. With the discovery of anticancer peptides, great progress has been made in cancer treatment. For the purpose of prompting the application of anticancer peptides in cancer treatment, it is necessary to use computational methods to identify anticancer peptides (ACPs). In this paper, we propose a sequence-based model for identifying ACPs (SAP). In our proposed SAP, the peptide is represented by 400D features or 400D features with g-gap dipeptide features, and then the unrelated features are pruned using the maximum relevance-maximum distance method. The experimental results demonstrate that our model performs better than some existing methods. Furthermore, our model has also been extended to other classifiers, and the performance is stable compared with some state-of-the-art works.

Highlights

Cancer is a serious health issue worldwide [1,2], and millions of people die of it.Traditional treatment methods focus on killing cancer cells, but at the same time normal cells are killed and there are high costs involved [3,4]
Our model has been extended to other classifiers, and the performance is stable compared with some state-of-the-art works
The Mathews correlation coefficient (MCC) value for our method is 0.8301, and the MCC value of iACP is 0.8058. iACP is a predictor based on the SVM, and the peptide is represented by g-gap dipeptide model

Summary

Introduction

Traditional treatment methods focus on killing cancer cells, but at the same time normal cells are killed and there are high costs involved [3,4]. This situation has changed with the discovery of anticancer peptides (ACPs). Because ACPs can interact with the anionic cell membrane components of cancer cells, cancer cells can be killed selectively by the ACPs without impairing the normal cells [5,6]. Anticancer peptides do not impair the body’s physiological functions, providing a new direction for cancer treatment. Treatment methods involving anticancer peptides have been receiving increasing attention. ACPs are represented by short peptides with 5 to 30 amino acids

Methods

Discussion

Conclusion