NAFLDkb

NAFLDkb provides a curated knowledge base and computational platform for computer-aided drug design targeting nonalcoholic fatty liver disease (NAFLD) by integrating multiperspective data to support identification and generation of therapeutic candidates.

Key Features:

• Knowledge graph construction: Curates multiperspective information from source materials and public databases to create comprehensive knowledge graphs linking drug-related entities and NAFLD.
• Chemical structure search: Implements chemical structure search capabilities for querying compound relationships.
• Drug-likeness screening: Provides drug-likeness screening functions to evaluate candidate compounds.
• Knowledge-based repositioning strategies: Applies knowledge-based repositioning strategies and a knowledge graph repositioning model that identified three novel drug candidates.
• Generative neural network model: Includes a generative neural network model that generated 137 new lead-like compounds.
• Research article annotation: Offers research article annotation features to extract and link literature-derived associations.
• Machine learning integration: Integrates existing data records and machine learning tools to support discovery workflows.

Scientific Applications:

• Computer-aided drug design for NAFLD: Supports identification and generation of candidate therapeutics for nonalcoholic fatty liver disease.
• Drug repositioning: Enables knowledge-based drug repurposing and discovery of novel repurposing candidates (three identified in case studies).
• Lead generation: Facilitates de novo lead-like compound generation via a generative neural network model (137 compounds produced).
• Knowledge extraction and association mapping: Elucidates associations between drug-related entities and NAFLD through knowledge graphs and literature annotation.

Methodology:

Curates multiperspective data from source materials and public databases; constructs comprehensive knowledge graphs; performs chemical structure search and drug-likeness screening; annotates research articles; and applies a knowledge graph repositioning model and a generative neural network model, integrating existing data records and machine learning tools.