Predicting Drug–Target Interactions from Drug Structure and Protein Sequence Using Novel Convolutional Neural Networks: A Review

Abstract

Drug–target interaction prediction is a cornerstone of computational drug discovery, therefore allowing the identification of putative therapeutic targets and accelerating drug repositioning. Experimental identification of DTIs is costly, time-consuming, and in feasible at large scale. Recent deep learning advances, especially convolutional neural Net works, or CNNs, have considerably enhanced DTI prediction by learning complicated hierarchical patterns directly from drug molecular structures and protein sequences. This Review provides a comprehensive overview of CNN-based prediction methods for DTI, covering representation learning strategies, network architectures, benchmark datasets, evaluation metrics, comparative performance analysis, challenges, and future research di rections. This study highlights the superiority of CNN models over traditional machine learning approaches and discusses emerging trends such as attention mechanisms, graph neural networks, and multi-modal data integration.