Improving Drug-Target Affinity Prediction Using Dynamic Graph Attention Network with Multi-scale Features and Attention Mechanism

Drug-target affinity (DTA) prediction is essential in drug discovery because traditional methods are timeconsuming and expensive.Yet, recent computational approaches often struggle with limitations in representing the structural and sequential complexities of drugs and proteins, resulting in inferior prediction performance.Therefore, this study proposes a method to enhance DTA prediction accuracy using Dynamic Graph Attention Networks (GATv2) and Bidirectional Long Short-Term Memory (BiLSTM).The model incorporates multi-scale features, which include drug motif graphs, and a three-way multi-head attention mechanism to capture complex interactions between drug and protein representations.Evaluated on Davis and KIBA datasets, the proposed model outperformed baseline models (e.g., GCN, GAT, 1DCNN, LSTM) and benchmark methods (e.g., GraphDTA, MSGNN-DTA, and DGDTA) across three evaluation metrics, achieving MSE of 0.3209 and 0.1864, CI of 0.8646 and 0.8616, and r2m of 0.5046 and 0.6672, respectively.This approach addresses limitations in static attention mechanisms, lack of multi-scale representation, and simplified interaction modeling in existing methods, offering a more robust method for DTA prediction.