Effective drug-target interaction prediction with mutual interaction neural network.

Area of Science:

• Computational chemistry
• Bioinformatics
• Drug discovery

Background:

• Accurate drug-target interaction (DTI) prediction is vital for efficient drug discovery.
• Deep learning models have advanced DTI prediction but face challenges in representing drugs and targets.
• Existing methods struggle to model the mutual influence between drugs and targets.

Purpose of the Study:

• To propose MINN-DTI, a novel deep learning model for enhanced DTI prediction.
• To jointly utilize molecular graphs and target distance maps for drug and target representation.
• To effectively capture the two-way impact between drugs and targets.

Main Methods:

• Developed MINN-DTI, integrating an Interformer module and an improved Communicative Message Passing Neural Network (CMPNN) called Inter-CMPNN.
• Represented drugs using molecular graphs and targets using distance maps.
• Modeled the mutual impact between drugs and targets for improved interaction prediction.

Main Results:

• MINN-DTI demonstrated superior performance compared to state-of-the-art methods on DUD-E, human, and BindingDB datasets.
• The model achieved better prediction accuracy for drug-target interactions.
• MINN-DTI offers interpretability by highlighting key amino acids and atoms in drug-target interactions.

Conclusions:

• MINN-DTI provides a powerful and interpretable approach for DTI prediction.
• The model's ability to capture mutual drug-target impacts advances the field.
• The proposed method has significant implications for accelerating drug discovery pipelines.