Background
In 1986, Don R. Swanson, an information scientist, discovered that medical literature is siloed — two fields may both publish about a shared biological mechanism (A) without ever citing each other. If disease B is linked to A, and compound C also affects A, then C may treat B — even though no paper has ever studied them together. This tool applies that insight at scale across curated biomedical databases to surface genuine leads that could accelerate drug discovery.
Applications
- –Drug repurposing — approved compounds that may treat diseases never tested against
- –Target discovery — shared biological mechanisms between a disease and a potential therapy
- –Hypothesis generation — candidates worth investigating in the lab or clinic
- –Literature gaps — connections that exist in data but haven't been studied together
Data Sources
- B↔ACTDComparative Toxicogenomics Database — manually curated disease–gene associations backed by direct experimental evidence., Open TargetsGenetic, somatic, and genomic evidence linking diseases to target genes from the Open Targets Platform., DISEASESJensen Lab — curated and high-confidence text-mined disease–gene associations. CC BY 4.0. — disease–gene associations
- C↔ACTDComparative Toxicogenomics Database — human curated chemical–gene interactions., DGIdbDrug–Gene Interaction Database — aggregates interactions from DrugBank, PharmGKB, ChEMBL, and ~30 other sources., DrugCentralClinically active drug–target pairs curated from FDA labels, WHO essential medicines, and biomedical literature. — drug–gene interactions
- B↔CPubMedA local co-occurrence index built from NCBI's MEDLINE archive (83.5M+ citations) — checks whether the disease and compound are already indexed together in the published literature. — checks if a connection already exists in print
- B↔C confirmedCTDCTD's direct chemical–disease curation — a human curator's own therapeutic or mechanism/marker verdict, bypassing the bridge gene entirely. Covers about 8% of pairs; shown as a Confirmed badge. — direct curated evidence for a minority of pairs
How to Use
- 01Filter by column — click any column label to open a text filter. Type a term, press Enter or comma.
- 02Sort — click the arrow next to any column label. Click the score label to switch metric.
- 03Source filter — click the Sources column header to filter by supporting database.
- 04Rx only — toggle to show only hypotheses where the compound (C) is an FDA-approved drug.
- 05Confirmed filter — show only, or hide, hypotheses with direct CTD compound–disease evidence (the ✓ Confirmed badge).
- 06Withdrawn filter — show only, or hide, compounds ChEMBL flags as withdrawn or no longer marketed.
- 07AI Assessment — open any hypothesis and click “Assess with AI” for an automated second opinion on whether the compound plausibly helps or harms the disease, grounded in the evidence already gathered. Not a validated finding.
- 08Star & export — click ☆ on any row, then Export CSV to download picks.