Target Discovery Comparison
Open Targets vs DisGeNET vs PharmGKB: Target Discovery Databases (2026)
Last updated: 2026-04-17
Target identification and validation is the first critical decision in drug discovery — pick the wrong target and years of effort are wasted. Open Targets (Wellcome/EMBL-EBI/Sanger), DisGeNET (MedBioInformatics), and PharmGKB (Stanford) approach this problem from different angles. Open Targets integrates genetic evidence with drug and pathway data to score target-disease associations. DisGeNET aggregates gene-disease associations from curated repositories and text mining. PharmGKB focuses specifically on pharmacogenomics — how genetic variation affects drug response. Together they cover the full spectrum from target discovery to pharmacogenetic implementation.
Open Targets Platform
EBI / Genentech / GSK / MSD / Pfizer / Sanofi / Wellcome Sanger
DisGeNET
IMIM / DisGeNET (commercial entity)
Head-to-Head
Structured comparison across key dimensions.
| Dimension | Open Targets Platform | DisGeNET | |
|---|---|---|---|
| Primary focus | Target-disease association scoring for drug discovery prioritization | Comprehensive gene-disease association aggregation | Pharmacogenomics — genetic variants affecting drug response |
| Data sources | GWAS Catalog, UniProt, ChEMBL, Reactome, ClinVar, OMIM, Expression Atlas, SLAPenrich, PheWAS, cancer biomarkers | UniProt, ClinVar, GWAS Catalog, CTD, ORPHANET, PsyGeNET, LHGDN, BeFree text mining | Manual expert curation from published literature, FDA drug labels, CPIC/DPWG guidelines |
| Scoring system | 0-1 association score per data source + overall harmonic sum; target prioritization factors | GDA Score (0-1) combining number of sources, source type, and publication count; VDA scores for variants | Levels of evidence (1A-4) for clinical annotations; manually assigned by expert curators |
| Disease coverage | ~30,000 diseases/phenotypes (EFO ontology); strong on common disease GWAS | ~30,000 diseases (UMLS, DO, HPO, OMIM); broad coverage including rare diseases | ~1,000 drugs × clinical annotations; focused on drugs with pharmacogenomic evidence |
| API / programmatic access | GraphQL API; BigQuery data dumps; comprehensive and well-documented | REST API (free tier + paid PLUS/Academic); RDF/SPARQL endpoint; R package | REST API; downloadable TSV datasets; limited query flexibility |
| Unique strength | Genetics portal with colocalisation analysis; L2G (locus-to-gene) scoring; target tractability assessment | Largest single collection of gene-disease associations (~1.1M GDAs); integrates curated + text-mined evidence | Gold-standard pharmacogenomic annotations; CPIC dosing guidelines; FDA label annotations |
| Web interface | Polished — disease/target/drug pages with evidence visualization and target prioritization view | Functional — searchable by gene, disease, or variant; network visualization | Specialized — drug/gene/variant pages with clinical annotations, pathways, and dosing guidelines |
| Update frequency | Quarterly releases (e.g., 24.12); automated pipeline | Periodic updates (~yearly); DisGeNET PLUS updated more frequently | Continuous curation; guidelines updated as evidence emerges |
| License | Open access (CC0 for data); fully free | Free for academic (CC BY-NC-SA); commercial requires DisGeNET PLUS license | Free for research (CC BY-SA-NC); commercial use requires license agreement |
| Key limitation | Biased toward common disease GWAS; less coverage of rare mendelian diseases | Text-mined associations can be noisy; free version update lag vs PLUS | Narrow scope (pharmacogenomics only); not a general target discovery platform |
When to Use Each
Open Targets Platform
You're prioritizing drug targets for a disease indication. You want scored associations integrating GWAS, OMIM, ChEMBL, and pathway data. You need an API for programmatic access and integration into target selection pipelines.
DisGeNET
You need the broadest collection of gene-disease associations across multiple evidence sources. You're doing network-based target discovery. You want to compare curated vs text-mined evidence for a given gene-disease pair.
Practitioner Verdict
Use Open Targets for systematic target prioritization with integrated genetic, drug, and pathway evidence — it's the most comprehensive platform for early target identification. Use DisGeNET when you need the broadest coverage of gene-disease associations from multiple curated sources. Use PharmGKB when your focus is pharmacogenomics — understanding how genetic variants affect drug response and dosing.
Stay updated on these tools
Weekly briefing on AI tool releases, benchmarks, and what works in drug discovery.