DRPMKB1.0: A Comprehensive Knowledge Base for an AI-Oriented Drug Repositioning Prediction Model

Area of Science:

• Computational biology
• Artificial intelligence in drug discovery
• Pharmacology

Background:

• Drug repositioning (DR) accelerates drug development by finding new uses for existing drugs.
• The proliferation of artificial intelligence (AI) models in DR necessitates effective integration and resource management.
• Large language models (LLMs) offer broad applicability but benefit from personalized knowledge bases for improved accuracy.

Purpose of the Study:

• To develop a comprehensive, AI-oriented knowledge base for drug repositioning prediction (DRPMKB 1.0).
• To create a standardized framework for evaluating and integrating diverse AI models and datasets for DR.
• To enhance the accuracy and efficiency of personalized drug repositioning through tailored model recommendations.

Main Methods:

• Compiled data from PubMed up to March 2024, covering 45 categories, 193 models, and 693 data entries.
• Developed DRPMKB 1.0 with display and interaction interfaces across data, model, application, and reference dimensions.
• Established a dual-evaluation framework to assess inherent model quality and predictive evidence.

Main Results:

• DRPMKB 1.0 provides a centralized data-sharing platform for DR.
• The dual-evaluation framework standardizes model selection and appraisal.
• Personalized recommendations based on user data significantly improve DR prediction accuracy.

Conclusions:

• DRPMKB 1.0 offers a robust, integrated platform for AI-driven drug repositioning.
• The knowledge base facilitates seamless integration of diverse data and models, supporting continuous AI enhancement.
• This resource empowers researchers with tailored model recommendations, advancing personalized drug discovery efforts.