drug-discovery

Query ChEMBL, PubChem, OpenFDA, and OpenTargets APIs for compound bioactivity and drug-likeness analysis.

Updated May 18, 2026
One-click install
npx skills add https://github.com/ossoolli/Nexum-Core --skill drug-discovery-ossoolli
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: drug-discovery
Source: https://github.com/ossoolli/Nexum-Core/tree/main/optional-skills/research/drug-discovery
Command: npx skills add https://github.com/ossoolli/Nexum-Core --skill drug-discovery-ossoolli

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? Pharmaceutical research requires pulling compound data, bioactivity measurements, and safety information from multiple public databases, each with its own API and query syntax. This Skill consolidates those lookups into ready-to-run workflows so researchers can screen molecules and assess drug-likeness without writing API client code. ## Core Features & Use Cases - Bioactive Compound Search: Query ChEMBL by target name or ID to retrieve active compounds with pChEMBL potency values. - Drug-Likeness Screening: Calculate Lipinski Rule of Five, Veber rules, TPSA, and QED scores for any compound via the PubChem API, including batch screening of multiple molecules. - Safety & Interaction Lookup: Retrieve drug-drug interactions and reported adverse events from OpenFDA labels, plus gene-disease associations from OpenTargets. - Use Case: A medicinal chemist evaluating a lead series can batch-screen candidate molecules against Ro5 and Veber criteria, then check the top hit for known FDA-reported interactions in one session. ## Quick Start Ask the assistant to check whether aspirin and ibuprofen pass the Lipinski Rule of Five and look up their known drug interactions.

Frequently Asked Questions about drug-discovery

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I search ChEMBL for compounds active against a target?▼

Query the ChEMBL target search endpoint with a target name like EGFR to get its ChEMBL ID, then fetch activities filtered by pChEMBL value. The included chembl_target.py script automates this and lists unique molecules ranked by potency.

How to check Lipinski Rule of Five for a compound without RDKit?▼

Fetch MolecularWeight, XLogP, H-bond donor, and acceptor counts from the PubChem PUG REST API, then apply the Ro5 thresholds in Python. The ro5_screen.py script does this in batch mode using only the standard library.

Does this require API keys for ChEMBL or PubChem?▼

No. All APIs used—ChEMBL, PubChem, OpenFDA, and OpenTargets—are free, public, and require no authentication. Only curl and python3 are needed as prerequisites, with a suggested 1-second delay between batch ChEMBL requests.

Can I look up drug-drug interactions from FDA data?▼

Yes. Query the OpenFDA drug label endpoint with the drug_interactions field to retrieve interaction text from official FDA labels. A separate endpoint aggregates reported adverse events, though these reflect reports rather than proven causation.

What are the limitations of API-based drug-likeness screening?▼

PubChem property lookups depend on the compound existing in the database by name, so novel structures cannot be screened this way. Predictions are rule-based estimates and should be combined with experimental ADMET data for real decisions.