CAHNetF-DTP: A Community-Aware Heterogeneous Network-Based Embedding Framework for Drug-Target Interaction Prediction

Area of Science:

• Bioinformatics
• Computational Biology
• Drug Discovery

Background:

• Drug-target interactions (DTIs) are crucial for drug discovery, but experimental prediction is costly and time-consuming.
• Existing network-based DTI prediction methods struggle to capture complex topological and biological relationships.
• Identifying hidden patterns in drug-target networks is essential for efficient drug development.

Purpose of the Study:

• To propose a novel computational framework, CAHNetF-DTP, for accurate drug-target interaction prediction.
• To leverage a heterogeneous biomedical network and community-aware strategies for enhanced DTI prediction.
• To address limitations in existing methods by capturing deeper biological and topological similarities.

Main Methods:

• Developed CAHNetF-DTP, a community-aware heterogeneous network-based framework for DTI prediction.
• Integrated 15 similarity-based subnetworks from drugs, targets, diseases, and side effects.
• Utilized Word2Vec for entity embedding generation and a controlled negative sampling strategy for class imbalance.

Main Results:

• CAHNetF-DTP demonstrated robust and competitive performance on benchmark datasets (KIBA, DAVIS) and real-world scenarios.
• The proposed method outperformed most existing DTI prediction tools, including SAR-based approaches.
• Case studies confirmed the model's capability in identifying novel drug-target pairs.

Conclusions:

• Integrating heterogeneous biological data through a community-aware network framework significantly improves DTI prediction.
• CAHNetF-DTP offers a valuable computational tool for accelerating drug discovery and identifying new therapeutic targets.
• The framework's ability to capture complex relationships highlights the potential of advanced network analysis in bioinformatics.