Book

Virus–host protein interactions play an important role in understanding how viruses interact with host organisms and how these interactions can be investigated computationally. Predicting such interactions on a large scale presents challenges related to data representation, computational efficiency, and the development of reliable predictive models. This book presents an optimized deep knowledge distillation framework for large-scale virus–host protein interaction prediction. It focuses on combining machine learning and deep learning techniques to build an efficient computational approach for classifying protein pairs. The framework integrates feature preparation, data balancing, feature representation learning, and knowledge distillation to address key challenges in predictive modeling. The presented approach explores the use of Gene Ontology information to enrich protein representations and autoencoder-based dimensionality reduction to obtain compact feature representations. It also incorporates data sampling strategies to construct balanced datasets and knowledge distillation to transfer predictive information from a larger teacher model to a more compact student model. Model performance is examined using established classification metrics, including accuracy, precision, recall, F1-score, Matthews correlation coefficient, and area under the receiver operating characteristic curve. The framework is designed to balance predictive performance with computational efficiency, making it relevant to researchers interested in scalable computational methods for protein interaction prediction. By bringing together knowledge distillation, feature learning, and machine learning-based classification, this book offers a methodological reference for researchers, postgraduate students, and practitioners working in artificial intelligence, computational biology, bioinformatics, and protein–protein interaction prediction. It also provides a foundation for further investigation into efficient deep learning frameworks for large-scale biological data analysis.