Drug-Target Prediction Based on Dynamic Heterogeneous Graph Convolutional Network.

Area of Science:

• Computational biology
• Bioinformatics
• Drug discovery

Background:

• Drug-target interaction (DTI) prediction is vital for drug discovery and repositioning.
• Graph neural networks (GNNs) show promise for DTI prediction but can lose information in sparse networks due to fixed thresholds.
• Dynamic Heterogeneous Graph (DT-DHG) addresses these limitations by optimizing information extraction.

Purpose of the Study:

• To propose a novel DTI prediction model, Dynamic Heterogeneous Graph (DT-DHG), that overcomes information loss in sparse networks.
• To enhance the utilization of insufficient information in drug-target interaction prediction.
• To improve the robustness and performance of GNN-based DTI prediction methods.

Main Methods:

• Developed a Dynamic Heterogeneous Graph (DT-DHG) model for DTI prediction.
• Incorporated progressive learning to dynamically adjust the receptive fields of nodes.
• Utilized DT-DHG to construct graphs without relying on empirically selected thresholds.

Main Results:

• The proposed DT-DHG model significantly improves upon existing GNN performance for DTI prediction.
• The method demonstrates robustness across different GNN backbones.
• DT-DHG outperforms current state-of-the-art methods in predicting novel drug-target interactions.

Conclusions:

• Dynamic Heterogeneous Graph (DT-DHG) effectively addresses information loss in sparse networks for DTI prediction.
• The model offers a robust and superior alternative to existing GNN-based approaches.
• DT-DHG shows significant potential for advancing drug discovery and repositioning through accurate novel DTI prediction.