Graph neural network approaches for drug-target interactions

Area of Science:

• Computational chemistry
• Bioinformatics
• Drug discovery

Background:

• Drug development is expensive, slow, and faces safety challenges.
• Predicting drug-target interactions (DTIs) is crucial for efficient drug discovery.
• Graph-based data, like molecular structures and protein networks, are suitable for graph neural networks (GNNs).

Purpose of the Study:

• To review deep neural networks for DTI prediction.
• To summarize databases used in DTI prediction.
• To introduce GNN applications in DTI prediction and discuss challenges.

Main Methods:

• Overview of deep neural networks in DTI models.
• Summary of essential databases for DTI prediction.
• Comprehensive introduction to GNN applications in DTI prediction.

Main Results:

• GNNs are effective in predicting DTIs.
• GNNs facilitate drug repositioning and accelerate drug discovery.
• The review highlights current challenges and future research avenues.

Conclusions:

• GNNs show significant promise in revolutionizing drug discovery.
• Further research in GNNs can overcome existing challenges in DTI prediction.
• This field holds potential for faster and safer drug development.