ArkDTA: attention regularization guided by non-covalent interactions for explainable drug-target binding affinity...

Area of Science:

• Computational chemistry
• Drug discovery
• Bioinformatics

Background:

• Protein-ligand binding affinity prediction is crucial for drug design.
• Deep learning models with cross-modal attention mechanisms offer improved explainability.
• Incorporating domain knowledge, such as non-covalent interactions (NCIs), is vital for enhancing these models.

Purpose of the Study:

• To propose ArkDTA, a novel deep neural architecture for explainable binding affinity prediction.
• To integrate NCIs into the protein-ligand attention mechanism for greater interpretability.
• To develop a more domain-aware deep drug-target interaction model.

Main Methods:

• Developed ArkDTA, a novel deep neural architecture.
• Integrated a cross-modal attention mechanism guided by NCIs.
• Utilized experimental data for model training and validation.

Main Results:

• ArkDTA achieved predictive performance comparable to state-of-the-art models.
• The model demonstrated significantly improved explainability.
• Qualitative analysis showed ArkDTA identifies potential NCI regions and guides model operations interpretably.

Conclusions:

• ArkDTA provides an explainable approach to binding affinity prediction.
• The NCI-guided attention mechanism enhances model interpretability and domain awareness.
• ArkDTA represents a significant advancement in explainable deep learning for drug-target interactions.