MLACP: machine-learning-based prediction of anticancer peptides

• DOI: 10.18632/oncotarget.20365
• Publication Date: 2017-08-19T11:56:45Z
• Authors: Balachandran Mana...; Shaherin Basith; Tae Hwan Shin; Sun Choi; Myeong Ok Kim; Gwang Lee
• Keywords: 0303 health sciences; 03 medical and health sciences; Research Paper; 3. Good health

Cancer is the second leading cause of death globally, and use therapeutic peptides to target kill cancer cells has received considerable attention in recent years. Identification anticancer (ACPs) through wet-lab experimentation expensive often time consuming; therefore, development an efficient computational method essential identify potential ACP candidates prior vitro experimentation. In this study, we developed support vector machine- random forest-based machine-learning methods for prediction ACPs using features calculated from amino acid sequence, including composition, dipeptide atomic physicochemical properties. We trained our Tyagi-B dataset determined machine parameters by 10-fold cross-validation. Furthermore, evaluated performance on two benchmarking datasets, with results showing that outperformed existing average accuracy Matthews correlation coefficient value 88.7% 0.78, respectively. To assist scientific community, also a publicly accessible web server at www.thegleelab.org/MLACP.html.