DrugR+

DrugR+ integrates data from DrugBank and KEGG with supervised machine learning datasets to identify drug repurposing, combination therapy, and replacement therapy candidates by predicting drug–target interactions.

Key Features:

• Database integration: Integrates data from DrugBank and KEGG to support analyses for drug repurposing and combination therapy.
• Single and synthetic repositioning: Supports analysis of both single drug repositioning and synthetic drug repositioning.
• Machine learning datasets: Provides four new datasets specifically for predicting drug–target interactions using supervised machine learning methods.
• Comparative ML analysis and case studies: Includes case studies and a comparative analysis of various machine learning methods applied to the generated datasets.
• Repurposing service: Implements a repurposing service that accepts a drug as input and outputs a list of candidate drugs for alternative applications.
• Data structure: Organizes information in normalized relational tables relevant to drug repurposing and combination applications.
• SQL query capabilities: Supports SQL queries for advanced data retrieval and analysis.
• Parallel processing updates: Supports updates and scalable processing using a map-reduce parallel processing method.

Scientific Applications:

• Drug repurposing: Identify potential new therapeutic uses for existing drugs via predicted drug–target interactions and dataset analyses.
• Combination therapy design: Prioritize candidate drug combinations and synthetic repositioning strategies for combination therapy research.
• Replacement therapy identification: Suggest alternative drugs for replacement therapy through repurposing predictions and database queries.
• Machine learning benchmarking: Benchmark and compare supervised machine learning methods for drug–target interaction prediction using the provided datasets and case studies.

Methodology:

Integrates DrugBank and KEGG data; constructs four new datasets; applies supervised machine learning to predict drug–target interactions; performs comparative analyses of various machine learning methods; stores data in normalized relational tables with SQL query support and uses map-reduce parallel processing for updates; exposes a repurposing service that accepts a drug input and returns candidate drugs.