DrugBank

DrugBank
Name	DrugBank
Developer	University of Alberta, IMS
Released	2006
Genre	Biomedical database
License	Commercial and academic licenses

DrugBank is a curated online resource that integrates detailed chemical, pharmacological, and pharmaceutical data with comprehensive drug target information. It serves researchers in pharmacology, medicinal chemistry, bioinformatics, and systems biology by linking drug entries to molecular targets, pathways, and clinical information. The resource is maintained by academic teams and used by industry, regulatory agencies, and healthcare organizations for drug discovery, repurposing, and safety assessment.

Overview

DrugBank combines structured drug records with annotated target and pathway relationships, providing cross-references to external resources such as PubChem, ChEMBL, KEGG, UniProt, and ClinicalTrials.gov. Entries typically include chemical structures, pharmaceutical formulations, pharmacokinetics, mechanism of action, and adverse effect profiles, mapping to protein targets annotated with identifiers from RefSeq, PDB, and Entrez Gene. The database supports computational tasks by offering downloadable data in formats compatible with tools like BLAST, Cytoscape, RDKit, and OpenBabel.

History and Development

Initially developed in the early 2000s by researchers at the University of Alberta and collaborators in Genome Alberta, the project built upon prior work in cheminformatics and proteomics exemplified by groups at Scripps Research, Broad Institute, and European Bioinformatics Institute. Major milestones include the 2006 public release, subsequent expansions to include experimental drug metabolites and comprehensive pharmacogenomic annotations linking to PharmGKB and dbSNP. Funding and partnerships have involved agencies and organizations such as the Natural Sciences and Engineering Research Council of Canada, academic consortia, and commercial biotechnology firms. Over time the platform incorporated crowd‑sourced literature curation practices used in resources like PubMed annotation projects and integrated structural data standards influenced by the Protein Data Bank community.

Content and Data Model

The content model organizes information into drug-centric and target-centric entities, representing small molecules, biologics, nutraceuticals, and experimental compounds. Chemical descriptors include SMILES and InChI strings cross-referenced to InChITrust standards, molecular weight, and stereochemistry annotations consistent with practices at IUPAC and ChemSpider. Pharmacological attributes such as bioavailability, half‑life, and metabolism pathways reference enzymes like members of the Cytochrome P450 family and transporter families annotated against UniProt accessions. Target annotations link to protein structures in the Protein Data Bank and gene-level data in Ensembl and HGNC, while pathway contexts use mappings to Reactome and KEGG PATHWAY entries. Data provenance traces literature citations from journals such as The Lancet, New England Journal of Medicine, Nature Medicine, and regulatory labels from agencies like the Food and Drug Administration and the European Medicines Agency.

Access and Licensing

Access is provided via a web interface with programmatic options including RESTful endpoints and data dumps suited for bulk analysis, echoing service models used by NCBI and EMBL-EBI resources. Licensing tiers range from free academic access for noncommercial research to paid commercial subscriptions for proprietary use, similar to licensing schemes at Clarivate and Elsevier commercial databases. Institutional subscriptions and site licenses are common among pharmaceutical companies, university research centers, and government laboratories such as those at National Institutes of Health and national health agencies. The platform also implements user authentication, usage tracking, and data use agreements to comply with contractual and intellectual property requirements observed in collaborations with industry partners like Pfizer and GlaxoSmithKline.

Applications and Impact

Researchers use the database for target identification, in silico screening, and drug repurposing studies that reference methods from machine learning workflows developed at institutions like MIT and Stanford University. Drug safety scientists integrate DrugBank annotations with adverse reporting systems such as FAERS to prioritize signals and evaluate pharmacovigilance hypotheses. Structural biologists and medicinal chemists leverage cross-links to Protein Data Bank structures and cheminformatics toolchains in academic groups at Harvard University and University of California, San Francisco to design analogs and study binding modes. The resource has contributed to publications in high-impact journals and to open drug-repositioning initiatives partnered with consortia like Innovative Medicines Initiative.

Limitations and Criticism

Critiques focus on curation lag, incompleteness for novel biologics, and challenges in maintaining up-to-date regulatory labeling changes, similar to criticisms leveled at other large aggregators like DrugCentral and commercial compound libraries. The hybrid licensing model has drawn debate in academic communities represented by organizations like SPARC and Creative Commons advocates, who argue for more open access consistent with data-sharing policies at agencies such as the Wellcome Trust. Technical limitations include variable annotation granularity and occasional inconsistencies in cross-references compared with primary resources such as UniProt and PubChem, necessitating cautious validation when used in regulatory submissions or clinical decision support systems.

Category:Biological databases