Drug-drug interactions prediction based on deep learning and knowledge graph: A review

Area of Science:

• Pharmacology
• Bioinformatics
• Computational Biology

Background:

• Drug-drug interactions (DDIs) pose significant risks, leading to unpredictable effects and adverse events.
• Conventional methods for DDI detection are often costly and time-intensive.
• There is a critical need for efficient computational approaches to predict DDIs.

Purpose of the Study:

• To systematically review research on drug-drug interaction prediction utilizing deep learning and knowledge graph techniques.
• To summarize available biomedical data and public drug databases relevant to DDI prediction.
• To analyze and compare existing computational DDI prediction methods.

Main Methods:

• Review of deep learning-based DDI prediction methods.
• Review of knowledge graph-based DDI prediction methods.
• Analysis of hybrid methods combining deep learning and knowledge graphs.
• Comparison of prediction methods on benchmark datasets.

Main Results:

• Categorization of DDI prediction methods into three main classes: deep learning-based, knowledge graph-based, and hybrid approaches.
• Comprehensive analysis of drug data and prediction methodologies.
• Comparative performance evaluation of various DDI prediction techniques.

Conclusions:

• Deep learning and knowledge graph techniques show promise for advancing DDI prediction.
• Further research is needed to address challenges such as asymmetric and high-order DDI prediction.
• Computational DDI prediction offers a more efficient alternative to traditional methods.