tooluniverse-network-pharmacology

Provide a compound and disease name, and the pipeline resolves identifiers via OpenTargets and DrugBank, builds compound-target-disease and PPI networks from STRING and IntAct, calculates network proximity, and ranks candidates with a 0-100 Network Pharmacology Score.

The pipeline queries over 15 databases through 60+ ToolUniverse tools, including OpenTargets, ChEMBL, DrugBank, STRING, IntAct, DGIdb, CTD, Reactome, Enrichr, FAERS, FDA, gnomAD, PubMed, EuropePMC, and ClinicalTrials.gov.

The score sums five components: network proximity (0-35 points based on Z-score), clinical evidence (0-25), target-disease association (0-20), safety profile (0-10), and mechanism plausibility (0-10). Scores of 80-100 indicate Tier 1 high repurposing potential.

Yes, it profiles all drug targets from OpenTargets, DrugBank, and CTD, classifies primary versus off-targets, computes disease module coverage, and assesses selectivity using DGIdb druggability and OpenTargets tractability data.

The pipeline applies fallback strategies, such as trying ChEMBL or PubChem when OpenTargets compound lookup fails, or using disease synonyms and parent ontology terms. Disconnected networks and missing data are reported explicitly rather than hidden.

Use network pharmacology when you need mechanistic insight from PPI networks, proximity Z-scores, and pathway overlap between drug targets and disease genes. For simple repurposing without network analysis, a dedicated drug repurposing skill is more appropriate.