DrugBank: a comprehensive resource for in silico drug discovery and exploration

Analysis

Why This Paper Matters

DrugBank addresses a critical gap in drug discovery by integrating chemical and biological data into a single, freely accessible database. Before DrugBank, researchers had to consult multiple disparate sources for drug properties and target information, slowing down in silico screening and target identification. By combining >4100 drug entries with >14,000 protein sequences, DrugBank enables rapid hypothesis generation and validation, accelerating the early stages of drug development.

The database's design, with over 80 data fields per entry and hyperlinks to major resources like KEGG, PubChem, and Swiss-Prot, makes it a central hub for pharmaceutical research. Its support for text, sequence, chemical structure, and relational queries allows flexible exploration, catering to both chemists and biologists. This integration has made DrugBank a standard reference in the field, evidenced by its high citation count (4044).

Technical Contributions

• Comprehensive Data Integration: Merges drug chemical data (e.g., structure, properties) with target protein data (e.g., sequence, function) in a single record.
• Rich Annotation: Each DrugCard contains >80 fields, covering drug nomenclature, pharmacology, pharmacokinetics, and target details.
• External Linking: Hyperlinks to KEGG, PubChem, ChEBI, PDB, Swiss-Prot, and GenBank enable cross-referencing and data enrichment.
• Advanced Search: Supports text, sequence (BLAST), chemical structure (substructure/similarity), and relational queries, making it versatile for various in silico tasks.
• Broad Applicability: Designed for drug target discovery, docking, metabolism prediction, interaction prediction, and education.

Results

The paper reports that DrugBank contains >4100 drug entries, including >800 FDA-approved small molecule and biotech drugs, and >3200 experimental drugs. It links these to >14,000 protein or drug target sequences. The database is fully searchable and freely available online. While no specific performance metrics are provided, the database's utility is demonstrated by its widespread adoption and 4044 citations.

Significance

DrugBank has had a profound impact on computational drug discovery by providing a unified, high-quality resource that bridges chemistry and biology. It has enabled researchers to perform in silico target identification, drug repurposing, and interaction prediction without needing to manually integrate data from multiple sources. The database's design has influenced subsequent bioinformatics resources and remains a cornerstone for pharmaceutical AI and machine learning applications. Its continued use underscores the importance of curated, integrated databases in accelerating drug development.