MINDG: a drug-target interaction prediction method based on an integrated learning algorithm.

Area of Science:

• Computational chemistry
• Bioinformatics
• Drug discovery

Background:

• Drug-target interaction (DTI) prediction is crucial for targeted therapeutics.
• Current computational methods face challenges in capturing deep structural features and integrating diverse data types.

Purpose of the Study:

• To address limitations in existing DTI prediction methods.
• To propose a novel framework that integrates sequence and structural information for improved DTI prediction.

Main Methods:

• A Multi-view Integrated learning Network (MINDG) was developed.
• MINDG employs a mixed deep network for sequence feature extraction.
• A higher-order graph attention convolutional network captures structural features.
• A multi-view adaptive integrated decision module enhances prediction.

Main Results:

• MINDG was evaluated on two datasets.
• The proposed method demonstrated improved DTI prediction performance.
• MINDG outperformed state-of-the-art baseline methods.

Conclusions:

• MINDG offers a powerful approach for DTI prediction by effectively integrating multi-view features.
• The framework advances computational drug discovery by enhancing prediction accuracy.