Trace paths
between a drug and a disease
Drugs, genes and diseases, with the bindings and associations between them, sit in one knowledge graph. Pick a disease to see repurposing candidates with their score evidence genes, then pick a candidate to explore paths from the drug to the disease.
Who is this for?A research exploration tool for drug repurposing researchers, early-stage biotech and pharma discovery staff, and students.Show moreHide
It is built for researchers working on drug repurposing (finding whether an existing drug could treat another disease), early-stage discovery staff at biotech and pharmaceutical companies, and bioinformatics and pharmacy students.
Pick a disease to see existing drugs with recorded binding to genes linked to the disease. The score evidence genes show why each drug is a candidate. Exploration paths from the drug to the disease are for forming hypotheses.
This page is not for patients or caregivers deciding on medication, or for clinicians making prescribing decisions. Candidates are research hypotheses, not proven treatments.
The public data sources are listed one by one in the sources and licenses section below. Before using it for company work, check the terms of each source. See sources and licenses
Search by name or ID (DOID:1612). Drugs and diseases can also be found by English synonyms and Korean names (aspirin · T2D). Genes and other types are found by name only. Synonyms and Korean names exist only for drugs and diseases listed in Wikidata. Diseases and drugs cover what Hetionet v1.0 contains.
Examples
- Disease
- Drug
- Gene
- Exploration path
This page is not medical advice. Results are computed from connections in the graph and do not establish efficacy or safety.
This page is provided free of charge for research and learning. The data comes from Hetionet v1.0 and its sources. Side effects (SIDER) and drug–gene bindings supported only by DrugBank are licensed for non-commercial use only. Each source keeps its own license as listed below. See sources and licenses
Sources and licenses
As of 2026-10-02
License counts (restrictions) were measured on 2026-10-02. The source table was last updated on 2026-11-09 when ChEMBL indications and UniChem were added; their fetch dates are given on each row.
Any description not yet translated is quoted from ONTOFLOW in the original Korean.
- Hetionet v1.0CC0 1.0 (each integrated source keeps its own terms)Integration of the whole graph
- DrugBank VocabularyCC0 1.0Drug names and IDs (1,552)
- WikidataCC0 1.0Synonyms and Korean names for 1,493 drugs and 136 diseases (Korean names: 634 drugs, 101 diseases)
- DrugBank 4.2CC BY-NC 4.0Non-commercialDrug–gene bindings (BINDS)
- SIDER 4.1CC BY-NC-SA 4.0Non-commercialShare-alike5,734 side effects; 138,944 drug–side effect relations
- ChEMBLCC BY-SA 3.0Share-alikeDrug–gene bindings (BINDS); drug–indication clinical phases and trial IDs (drug_indication, 60,055 records, ChEMBL 37, fetched 2026-11-09)
- UniChemCC BY 4.0DrugBank–ChEMBL drug ID mapping (linking clinical evidence, fetched 2026-11-09)
- DrugCentralCC BY 4.0Drug–gene bindings; 345 pharmacologic classes; 1,029 class–drug relations
- BindingDBCC BY 3.0Drug–gene bindings (BINDS)
- DisGeNET (as included in Hetionet)ODbL 1.0Share-alikeDisease–gene associations (ASSOCIATED_WITH)
- DISEASES · GWAS CatalogCC BY 4.0Disease–gene associations (ASSOCIATED_WITH)
- Disease Ontology (as included in Hetionet)CC BY 3.0137 diseases
- Disease Ontology v2026-10-29CC0 1.0Disease–side effect mapping (UMLS cross-references, 83 diseases); disease–clinical indication mapping (MeSH cross-references, 121 diseases)
- Reactome · WikiPathways · PIDCC BY 4.0 · CC BY 3.0 · CC01,822 pathways; 84,372 gene–pathway relations
- Entrez Gene · MeSH · MEDLINE · PharmacotherapyDB · HGNCCC0 / public domainGenes, symptoms, treats, palliates, presents, disease similarity, compound similarity, protein–gene
Data with conditions
Non-commercial
Drug–side effect relations and side effect names138,944 records
Supported by SIDER 4.1 only. CC BY-NC-SA 4.0
Drug–gene bindings supported only by DrugBank7,861 records
Out of 11,571 drug–gene bindings. 3,524 are also supported by ChEMBL, DrugCentral or BindingDB
No license
Drug–gene bindings supported only by unlicensed sources186 records
PDSP Ki · PubChem · US patents
Disease–gene associations supported only by DOAF632 records
Out of 12,623 disease–gene associations. 11,991 are supported by DISEASES, DisGeNET or GWAS
Source being confirmed
Compound–protein bindings, PubChem compounds, proteins28,240 records
The source dataset has not been confirmed yet. The format linking compounds (PubChem CID) and proteins (UniProt) suggests the BindingDB family
How it was counted Counted by querying the per-edge and per-node license property on the public Hetionet Neo4j (neo4j.het.io). A relation with several supporting sources is counted as permissively supported if at least one source has a permissive license.
Graph loading and scoring run on ONTOFLOW.