MolTrans: Molecular Interaction Transformer for drug-target interaction prediction.

Area of Science:

• Computational chemistry
• Bioinformatics
• Drug discovery

Background:

• Drug-target interaction (DTI) prediction is crucial for in-silico drug discovery.
• Current methods face challenges in accuracy, interpretability, and utilizing vast unlabeled molecular data.

Purpose of the Study:

• To develop a novel deep learning model, MolTrans, for improved DTI prediction.
• To address limitations in existing molecular representation learning and data utilization.

Main Methods:

• MolTrans employs a knowledge-inspired sub-structural pattern mining algorithm.
• An augmented transformer encoder captures semantic relations from unlabeled biomedical data.
• Incorporates an interaction modeling module for enhanced DTI prediction.

Main Results:

• MolTrans demonstrates improved DTI prediction performance compared to state-of-the-art baselines.
• The model offers more accurate and interpretable DTI predictions.
• Effectively leverages large unlabeled molecular datasets.

Conclusions:

• MolTrans represents a significant advancement in DTI prediction.
• The approach enhances the efficiency and effectiveness of in-silico drug discovery.
• The model's code is publicly available for further research.