Cancer is a leading cause of mortality worldwide, and conventional cancer treatments can inflict considerable harm on vital organs. Anticancer peptides (ACPs), a class of small peptides produced by the immune system, possess several advantages in-cluding high specificity, low immunogenicity, minimal toxicity, and high tolerance under normal physiological conditions. Consequently, they hold promise as a viable alternative to traditional anticancer drugs.
Traditional methods of identifying anticancer peptides involve wet-lab experiments, but their small size and low abundance make this approach inefficient and costly. Machine learning-based peptide prediction methods have emerged as a cost-effective and efficient alternative for identifying anticancer peptides from large biological datasets. This study introduces the concept of ensemble to develop an algorithmic model capable of predicting anticancer peptides.
The ACPPfel algorithm achieved high accuracy of 98.53% and an AUC value of 0.9972 in the ACPfel dataset. this has led to performance improvements on other datasets as well.