A comprehensive review of neural network-based approaches for drug-target interaction prediction

Predicting Drug-Target Interactions (DTI) is vital for accelerating drug discovery and repurposing. This review assesses the efficacy of neural network-based methods, including Convolutional Neural Networks (CNNs), Graph Neural Networks (GNNs), and Transformer architectures utilizing self-attention mechanisms, in predicting DTI. Leveraging recent studies and benchmark datasets, such as DrugBank, neural networks outperform traditional methods, with hybrid architectures achieving an Area Under the Receiver-Operating Characteristic Curve (AUROC) of 0.979. These models enhance precision and reduce errors by integrating multimodal data, such as molecular graphs and protein sequences. However, challenges like high computational costs, limited interpretability, and data scarcity persist. Future advances may combine multimodal data with explainable Artificial Intelligence (AI) to improve DTI analysis and unlock new therapeutic opportunities. This study highlights the transformative potential of neural networks in computational drug discovery.