Enhanced Protein–Ligand Binding Affinity Prediction Espoused Temporal Channel Reconfiguration Multi-Graph Convolution Network

Predicting protein–ligand binding affinity is vital for drug discovery, but remains challenging due to complex spatial interactions. To overcome these complications, Enhanced Protein–Ligand Binding Affinity Prediction espoused Temporal Channel Reconfiguration Multi-Graph Convolution Network (EPLBAP-TCRMCN-YDSEOA) is proposed in this paper. The proposed EPLBAP-TCRMCN-YDSEOA is a novel framework combining a Temporal Channel Reconfiguration Multi-Graph Convolution Network (TCRMCN) with the Young’s Double Slit Experiment Optimization Algorithm (YDSEOA). The model integrates Event-triggered Consensus Kalman Filtering (ECKF)-based preprocessing, Membership Fusion Semi-supervised Fuzzy Clustering Algorithm (MFSFCA)-driven 3D segmentation, and Directional Lifting Wavelet Transform (DLWT) for feature extraction, enabling dynamic and spatially-aware prediction. TCRMCN captures multi-level molecular interactions, while YDSEOA efficiently optimizes network parameters. The experiments conducted on the MMDRPv1 dataset demonstrate superior performance, with the proposed method achieving up to 99.6% accuracy and outperforming existing approaches in terms of RMSE, MAE and Pearson correlation coefficient. These results highlight the model’s strong potential to support advancements in computational drug design.