skills.rest
K-Dense-AI
K-Dense-AIOFFICIAL
Explore all skills by this author →
Website𝕏@k_dense_aiGitHub

ncats-arax

Queries the NCATS ARAX API for bounded, provenance-rich biomedical knowledge-graph relationships.

41.1k|3.8k|Updated Oct 19, 2025
Search Tags:#provenance#biomedical#Knowledge Graph#ncats translator#trapi#arax#biolink
One-click install
npx skills add https://github.com/K-Dense-AI/scientific-agent-skills --skill ncats-arax
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ncats-arax
Source: https://github.com/K-Dense-AI/scientific-agent-skills/tree/main/skills/ncats-arax
Command: npx skills add https://github.com/K-Dense-AI/scientific-agent-skills --skill ncats-arax

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Biomedical researchers need trustworthy, provenance-rich relationships from the NCATS Translator knowledge graph, but raw TRAPI queries are complex, error-prone, and easy to over-interpret. This Skill constrains ARAX access to reviewed, typed one-hop and endpoint-pinned two-hop lookups with full provenance preservation.

Core Features & Use Cases

  • Bounded graph lookups: Run Biolink-typed one-hop and exactly two-hop queries against RTX-KG2 or two to five explicitly named federated providers, with qualifier constraints and a hard 50-result cap.
  • Entity normalization: Normalize free-text terms to canonical CURIEs and categories for review before any graph query is submitted.
  • Provenance inspection: Save exact request/response bytes with SHA-256 hashes, inspect TRAPI edge bindings, publications, and knowledge-source provenance, and rebuild bounded summaries offline.
  • Use Case: A researcher investigating whether imatinib decreases ABL1 activity runs a qualifier-aware one-hop query, then inspects the saved summary.json for edge bindings, primary sources, and publication IDs before verifying candidates in the literature.

Quick Start

Ask the agent to run a one-hop ARAX lookup for how a specific drug CURIE affects a specific gene CURIE, acknowledging the query is public and saving artifacts to a new output directory.

Frequently Asked Questions about ncats-arax

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

FAQPage Schema
How do I query the NCATS ARAX knowledge graph for drug-gene relationships?▼

Run the arax_client.py one-hop command with pinned subject and object CURIEs, Biolink categories, and a predicate such as biolink:affects. The client defaults to RTX-KG2 lookup mode with 20 results and saves request.json, response.json, summary.json, and manifest.json.

How do I convert a disease or drug name into a CURIE for a TRAPI query?▼

Use the normalize command with the free-text term and an expected Biolink category. It returns the canonical identifier, name, category, and a synonym preview for review; normalization never triggers a graph query automatically.

Does the ARAX client require an API key or external Python packages?▼

No API key is required and the client uses only the Python standard library. It requires Python 3.10+ and outbound HTTPS access to arax.transltr.io.

Can I run multi-hop pathfinding or inference queries with ARAX through this client?▼

No. The client deliberately supports only typed one-hop and endpoint-pinned two-hop lookups, excluding inference, ranking, Pathfinder, three-hop, and open-ended pathfinding. Those limits cannot be bypassed with direct HTTP calls under this skill.

What does exit code 7 mean when running a federated ARAX query?▼

Exit code 7 indicates a partial federated response where one or more knowledge providers timed out or errored, but useful results were retained. Artifacts are preserved and the run is marked partial in the manifest.

Is it safe to submit patient-specific or confidential questions to ARAX?▼

No. ARAX status facilities may expose query and caller metadata even with store=false, so only public, nonsensitive research questions should be submitted. Never include patient information, unpublished compound programs, or proprietary target hypotheses.

Related Skills

Explore More Education & Research Skills→
nahisahonahisaho
3

scientific-text-mining-nlp

Extract biomedical entities and relationships from scientific text into a literature knowledge graph.

Community
Advanced
mims-harvardmims-harvard
1.6k

drug-repurposing

Identify drug repurposing candidates across disease targets, drugs, and safety data.

Official
Advanced
markwang2658markwang2658
31

drug-discovery

Identify and analyze bioactive compounds using ChEMBL, PubChem, OpenFDA, and OpenTargets.

Community
Intermediate
zainhaszainhas
2

together-batch-inference

Batch-process inference requests asynchronously via Together AI's Batch API.

Community
Advanced