Heterogeneous network drug-target interaction prediction model based on graph wavelet transform and multi-level contrastive learning

TL;DR

The paper tackles the challenge of predicting drug-target interactions with interpretability by proposing a heterogeneous-network framework that fuses local neighbor information and deep cross-type associations through a graph wavelet transform (GWT) encoder and a graph neural network (HGCN) encoder. It introduces a cross-view, multi-scale contrastive learning scheme to align representations from the neighborhood and deep views, and applies DistMult with a reconstruction-based objective to predict interactions while enabling mechanism decoding. Key contributions include multi-scale feature extraction via the MG Encoder, heterogeneous data fusion with cross-graph attention, and physics-informed knowledge integration through docking-inspired signals. Empirical results show robust, high-performance predictions (AUC/AUPR) across datasets, with ablation analyses confirming the importance of each module for interpretability and accuracy, suggesting practical value for targeted drug discovery and mechanism elucidation.

Abstract

Drug-target interaction (DTI) prediction is a core task in drug development and precision medicine in the biomedical field. However, traditional machine learning methods generally have the black box problem, which makes it difficult to reveal the deep correlation between the model decision mechanism and the interaction pattern between biological molecules. This study proposes a heterogeneous network drug target interaction prediction framework, integrating graph neural network and multi scale signal processing technology to construct a model with both efficient prediction and multi level interpretability. Its technical breakthroughs are mainly reflected in the following three dimensions:Local global feature collaborative perception module. Based on heterogeneous graph convolutional neural network (HGCN), a multi order neighbor aggregation strategy is designed.Multi scale graph signal decomposition and biological interpretation module. A deep hierarchical node feature transform (GWT) architecture is proposed.Contrastive learning combining multi dimensional perspectives and hierarchical representations. By comparing the learning models, the node representations from the two perspectives of HGCN and GWT are aligned and fused, so that the model can integrate multi dimensional information and improve the prediction robustness. Experimental results show that our framework shows excellent prediction performance on all datasets. This study provides a complete solution for drug target discovery from black box prediction to mechanism decoding, and its methodology has important reference value for modeling complex biomolecular interaction systems.

Figures (5)

• Figure 1: Feature acquisition of four types of nodes using PCA and Node2vec
• Figure 2: GHCDTI overall model architecture
• Figure 3: Results of running four models on the dataset
• Figure 4: Distribution of attention weights assigned to four node perspectives
• Figure 5: Ablation experiment results after removing parts of the model