Structure-inclusive similarity based directed GNN: a method that can control information flow to predict drug-target...

Area of Science:

• Computational chemistry
• Drug discovery
• Bioinformatics

Background:

• Drug-target association is crucial for drug discovery and repurposing.
• Predicting binding affinity offers more detailed insights than binary classification.
• Existing methods often use symmetric similarity, neglecting substructure inclusion relationships.

Purpose of the Study:

• To propose a novel similarity measure that accounts for substructure inclusion.
• To develop a graph convolutional network-based model for drug-target binding affinity prediction using this new similarity.

Main Methods:

• Introduced Structure-Inclusive Similarity (SIS) to capture substructure inclusion between molecules.
• Constructed drug and target graphs based on SIS.
• Employed a graph convolutional network for binding affinity prediction.

Main Results:

• The SIS-based approach significantly improves the accuracy of drug-target binding affinity prediction.
• Our method demonstrates superior performance compared to several state-of-the-art techniques.
• Case studies confirm the practical utility of the model.

Conclusions:

• Considering substructure inclusion in similarity measures enhances prediction models.
• The proposed SIS method is effective for predicting drug-target binding affinity.
• This approach offers a valuable tool for drug discovery and repurposing efforts.