Akademik Veri Yönetim Sistemi
2023 7th International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT), Ankara, Türkiye, 26 - 28 Ekim 2023, ss.1-6
Anticancer peptides (ACPs) represent a class of peptidic compounds with promising antineoplastic properties. The utilization of ACPs in the realm of cancer prevention offers a potential alternative to conventional oncological interventions, primarily due to their heightened selectivity and improved safety profile. Recent advancements in scientific research have ignited enthusiasm for peptide-based therapeutic modalities, largely due to their exceptional discriminatory abilities. These therapeutic strategies offer a distinct advantage by precisely and effectively targeting malignant cells while preserving the integrity of normal cellular components. Nevertheless, the exponential growth in available peptide sequences has posed a significant challenge, the development of a robust and precise predictive framework. In this work, we present an effective model designed for the classification of anticancer peptides. For this purpose, support vector machines (SVMs), naive Bayes (NB), recurrent neural networks (RNNs), and gated recurrent unit (GRU) are evaluated. In an effort to elucidate the merits of our study, an extensive series of experiments has been systematically conducted. The datasets utilized in these investigations have been deliberately chosen from the well-established ENNACT, ENNAACT+A, ENNAACT+B, and ACP740 datasets. The experiment results show that the best classification accuracies are performed by GRU for the ENNACT dataset with 90.28%, by RNN with 91.44% for the ENNACT+A dataset, by RNN with 91.76% for ENNNACT+B dataset, and by GRU wit 97.12% for ACP740 dataset.