Graph-sequence attention and transformer for predicting drug-target affinity.

Area of Science:

• Computational chemistry
• Drug discovery
• Bioinformatics

Background:

• Drug-target binding affinity (DTA) prediction is crucial for efficient drug discovery.
• Traditional drug development is expensive, time-consuming, and carries safety risks.
• Drug repurposing offers a faster, safer alternative by identifying new uses for existing drugs.

Purpose of the Study:

• To develop an effective computational method for predicting DTA.
• To leverage attention mechanisms for improved prediction accuracy.
• To introduce a novel model, GSATDTA, for DTA prediction.

Main Methods:

• Utilized Bi-directional Gated Recurrent Units (BiGRU) and graph neural networks for drug representation.
• Employed an attention mechanism to fuse drug representations (SMILES and molecular graphs).
• Applied an efficient transformer to learn protein sequence representations, capturing long-range dependencies.

Main Results:

• The proposed GSATDTA model demonstrated superior performance compared to state-of-the-art methods.
• Extensive experiments validated the model's effectiveness on two independent datasets.
• The self-attention mechanism in GSATDTA effectively captures relevant features for DTA prediction.

Conclusions:

• GSATDTA offers a significant advancement in computational DTA prediction.
• The model's ability to integrate diverse molecular representations enhances prediction accuracy.
• GSATDTA shows promise for accelerating drug discovery and repurposing efforts.