ABSTRACT
The use of artificial intelligence (AI) in healthcare has the potential to revolutionize medical information processing and analysis. In particular, multimedia information processing is a critical area that can benefit from AI tools and techniques due to the vast number of medical images, videos, and audio recordings generated each day. One specific area of interest is the development of anticancer peptides (ACP) that are more potent and have fewer adverse effects than targeted treatment and chemotherapy for cancer medication. However, predicting ACP activity from genomic data is a difficult task in immunoinformatics. In this chapter, the work of an earlier researcher is discussed that presents the MLACP 2.0 (Machine Learning-based Anticancer Peptide Predictor) predictor, which utilizes machine learning–based methods to deduce ACP activity from protein peptide patterns. A variety of feature encodings were explored and trained using seven different conventional classifiers, which were then combined and trained using a convolutional neural network (CNN). MLACP 2.0 outperformed current ACP prediction tools and performed better than CNN-based embedding models and traditional single models during cross-validation and independent evaluation. The MLACP 2.0 predictor will simplify the identification of novel ACPs, facilitating the design of hypothesis-driven experiments and ultimately leading to improved patient outcomes in cancer treatment.
