gene-research

name	gene-research
description	Investigate genes, drug targets, treatments, disease associations, pharmacogenomics, and literature. Use when the user asks about gene symbols, ENSG IDs, targets, drug targets, drugs, therapies, disease links, PGx, or Open Targets.

Gene Research

Overview

Every gene query must use all three data sources:

Source	Tools	What it provides
Open Targets	lookup_gene, get_gene_disease_associations, search_gene_drugs, get_gene_drug_details, get_drug_details	Ensembl ID, disease links, related drugs with full profiles
ClinPGx local	get_gene_pgx_profile	PGx guidelines, allele effects, evidence levels
Literature	paper_search_search_pubmed (+ optional arXiv)	Paper metadata + abstracts

Do not call raw GraphQL MCP tools (query_open_targets_graphql is not available).

Mandatory workflow

Run all steps for every query. Do not pick a subset based on the question type.

1. Resolve gene identifiers (Open Targets)

Call lookup_gene with the gene symbol (e.g. BRCA2).

Returns ensembl_id (e.g. ENSG00000139618).

2. Query target-disease associations (Open Targets)

Call get_gene_disease_associations with the Ensembl ID from step 1.

Use limit=5 by default.

3. Query related drugs / treatments (Open Targets)

Call get_gene_drug_details with the gene symbol (limit=5 by default).

This returns clinical stage, indications, mechanisms of action, adverse events, gene-filtered pharmacogenomics, and PubMed references (title + DOI).

Use search_gene_drugs only when you need a lightweight drug name list. Use get_drug_details when you already have a specific ChEMBL ID.

Optional drug_hint for narrowing (e.g. "PARP inhibitor" for BRCA2).

4. Query pharmacogenomics profile (ClinPGx)

Call get_gene_pgx_profile with the gene symbol.

If the gene has no ClinPGx entry, note that and continue.

5. Literature evidence (PubMed)

Always call paper_search_search_pubmed (max_results=3).

Add paper_search_search_arxiv when the topic has computational or preprint angles.

Build the query from the gene symbol plus top disease/drug terms from steps 2–4 (e.g. "TPMT" AND (pharmacogenomics OR azathioprine)).

6. Summarize results

1. Gene: symbol, Ensembl ID
2. Disease associations: top diseases with scores
3. Drugs / treatments: names, ChEMBL IDs, clinical stage, indications, MOA, adverse events, Open Targets PGx, references
4. Pharmacogenomics: key drugs, guidelines, allele effects
5. Literature: titles, authors, year, URLs, brief abstract summary

Entity ID reference

Entity	ID format	Example
Target	ENSEMBL gene ID	ENSG00000139618 (BRCA2)
Disease	EFO / MONDO ID	MONDO_0007254 (breast cancer)
Drug	ChEMBL ID	CHEMBL521686 (olaparib)