Drug-target interaction (DTI) prediction plays a pivotal role in drug discovery. In recent years, deep learning-based models have been advanced rapidly, accelerating the identification of potential DTIs. However, how to effectively capture the cross-modal information from bidirectional DTIs and how to further fuse them remain challenges for existing methods. To address these issues, we propose a deep learning fusion framework termed cross-modal interaction-aware progressive fusion network (CIPFN) for DTI prediction. This framework introduces a bidirectional interaction-aware module to precisely align fine-grained interactions between drugs and proteins. In addition, a progressive fusion network is also developed, including both gated and convolutional fusion blocks, to efficiently extract critical information within drug-target relationships. Experimental results across five benchmark data sets demonstrate that the proposed CIPFN achieves significant improvements over some state-of-the-art methods on the metrics of AUROC, AUPRC, F1, sensitivity, and accuracy.
Cao et al. (Fri,) studied this question.