Multi-view based heterogeneous graph contrastive learning for drug-target interaction prediction.

Area of Science:

• Computational biology
• Bioinformatics
• Drug discovery

Background:

• Drug-Target Interaction (DTI) prediction is crucial for accelerating drug discovery.
• Existing methods often struggle with integrating topological and semantic information and preserving feature diversity.
• Challenges include insufficient information integration and diminished representation diversity in graph convolution operations.

Purpose of the Study:

• To propose a novel paradigm, Multi-view Based Heterogeneous Graph Contrastive Learning for Drug-Target Interaction Prediction (HGCML-DTI).
• To address the limitations of insufficient information integration and diminished representation diversity in DTI prediction.
• To enhance the expressiveness and discriminative ability of learned features for DTI prediction.

Main Methods:

• Constructing a drug-protein heterogeneous graph and deriving node representations using a weighted Graph Convolutional Network (GCN).
• Integrating topology and semantic graphs for Drug-Protein Pairs (DPP) into a unified public graph.
• Employing a multi-channel graph neural network and a multi-view contrastive learning strategy to learn DPP representations and preserve diversity.
• Utilizing a Multilayer Perceptron (MLP) for DTI recognition.

Main Results:

• The proposed HGCML-DTI method significantly outperforms seven competitive baselines across six real-world datasets.
• Demonstrated superior performance in Drug-Target Interaction (DTI) prediction tasks.
• Validated the effectiveness of combining multi-view learning and contrastive strategies.

Conclusions:

• The HGCML-DTI paradigm effectively addresses key challenges in DTI prediction by integrating diverse information sources.
• The study highlights the importance of multi-view learning and contrastive strategies for advancing DTI prediction.
• The proposed method offers a significant improvement over existing state-of-the-art approaches in drug discovery and development.