Adrecs

Skill BioTender-max/awesome-bio-agent-skills/skills/drugclaw/adrecs

Query the ADReCS (Adverse Drug Reaction Classification System) v3.3 database. Use whenever the user asks about adverse drug reactions, drug safety profiles, ADR classification, ADR severity/frequency, or wants to look up any entity (drug name, BADD Drug ID, DrugBank ID, ATC code, CAS RN, PubChem CID, KEGG ID, ADR term, ADReCS ID, MedDRA code, MeSH ID) in ADReCS.From its SKILL.md

Install

npx -y skills add BioTender-max/awesome-bio-agent-skills --skill adrecs

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.

SKILL.md

2.5 KB, 652 tokens by cl100k_base, as published. Nobody here has run it

ADReCS Query Skill

Search ADReCS v3.3 records by any entity. Auto-detects type by prefix:

Input Pattern	Detected As	Example
`BADD_D00142`	BADD Drug ID	exact on drug_id column
`DB00945`	DrugBank ID	resolved via Drug_information
`A02BC01`	ATC code	resolved via Drug_information
`50-78-2`	CAS RN	resolved via Drug_information
`CID2244` or bare digits	PubChem CID	resolved via Drug_information
`D00109` (5-digit)	KEGG ID	resolved via Drug_information
`08.06.02.001`	ADReCS ID	substring on ADReCS_ID column
`10003781` (8-digit)	MedDRA code	resolved via ADR_ontology
`D######` (6+ digit)	MeSH ID	resolved via ADR_ontology
anything else	free text	substring on drug_name OR ADR_term

API

Function	Input	Returns
`load_drug_adr(path)`	txt path	DataFrame (Drug–ADR pairs)
`load_drug_info(path)`	xlsx path	DataFrame (drug metadata)
`load_adr_ontology(path)`	xlsx path	DataFrame (ADR hierarchy)
`search(entity)`	single entity string	DataFrame of matching Drug–ADR rows
`search_batch(entities)`	list of entity strings	dict[str, DataFrame]
`summarize(hits, entity)`	DataFrame + label	compact LLM-readable text
`to_json(hits)`	DataFrame	list[dict]

Usage

See if __name__ == "__main__" block in 62_ADReCS.py for runnable examples covering: drug name lookup, BADD Drug ID, DrugBank ID, ADR term, ADReCS ID prefix, batch search, and JSON output.

Data

Source: ADReCS v3.3 — https://www.bio-add.org/ADReCS/
Primary file: Drug_ADR_v3.3.txt (drug–ADR association pairs)
Drug metadata: Drug_information_v3.3.xlsx (name, synonyms, DrugBank/KEGG/PubChem/ATC/CAS cross-refs)
ADR hierarchy: ADR_ontology_v3.3.xlsx (ADReCS ID, ADR term, MedDRA code, MeSH ID, 4-level classification)
Path: DATA_DIR variable in 62_ADReCS.py

Citation

Cai MC et al. (2015) ADReCS: an ontology database for aiding standardization and hierarchical classification of adverse drug reaction terms. Nucleic Acids Research 43(D1):D907–D913.

What ships with it: 5 files

20.1 KB alongside SKILL.md, 4 of them executable

adrecs_skill.pyruns3.3 KB
example.pyruns14.7 KB
__init__.pyruns130 B
README.md993 B
retrieve.pyruns1018 B