Gene¶
Use gene commands to retrieve canonical metadata and targeted biological context.
What the gene guide covers¶
- symbol-based retrieval,
- lightweight search,
- section expansion,
- JSON output for downstream systems.
Search genes¶
Start with search when you are unsure of symbol spelling or want to inspect alias candidates.
Known aliases that map to one canonical human gene can also be passed directly to get gene.
Useful fields in search output typically include symbol, Entrez ID, and species.
Get a gene record¶
The default gene view is concise and intended for orientation. When MyGene.info
returns genomic coordinates, BioMCP labels them with the GRCh38 genome build
(Chromosome (GRCh38): ...) so coordinate consumers do not have to infer the
reference. Its More: block keeps pathways, ontology, and diseases
visible and now also surfaces funding as a direct follow-up from the base
card.
Request deeper sections¶
BioMCP expands detail via positional sections.
Pathway view:
Disease associations:
Ontology terms:
Protein summary:
When UniProt exposes legacy protein names, the protein section includes an
Also known as: line with alternative full names and short names from UniProt.
When UniProt exposes alternative products, the same section also includes an
Isoforms (N) line with isoform names and the displayed isoform length when
that length is available from the base UniProt record.
GO terms and interactions:
CIViC evidence summary:
Tissue expression (GTEx):
Protein tissue expression and localization (Human Protein Atlas):
Druggability profile (DGIdb interactions plus OpenTargets tractability and safety):
Funding context (NIH Reporter grants mentioning the canonical symbol in the most recent 5 NIH fiscal years):
Diagnostic-test pivot (GTR tests for the gene):
The diagnostics and funding sections are opt-in and are not included in
biomcp get gene <symbol> all.
Gene-disease validity (ClinGen):
When ClinGen validity rows are present, the gene card's See also: block
promotes a recruiting-trial search keyed to the newest reviewed disease label
already shown on the card, ahead of the generic gene pivots.
Constraint metrics (gnomAD):
Multiple sections can be chained:
All supported sections:
To add funding, request it explicitly:
Helper commands¶
biomcp gene trials BRAF --limit 5
biomcp gene trials SHANK3 --limit 5
biomcp gene drugs BRAF --limit 5
biomcp gene pathways BRAF
biomcp gene articles BRAF
biomcp gene definition BRAF
For CTGov, supported rare-disease genes such as SHANK3 reuse the shared
trial plan so the search can include the associated condition expansion. Other
genes and non-CTGov sources keep the literal gene trial search.
Common workflows¶
Clinical trial pivot¶
Literature pivot¶
Variant pivot¶
Error handling expectations¶
If a section name is unsupported, BioMCP returns an explicit unknown-section message with hints about valid section names.
JSON mode¶
Use JSON for pipelines or agent post-processing.
biomcp --json get gene BRAF druggability includes DGIdb interaction fields plus
OpenTargets tractability[] modality summaries and safety_liabilities[] event summaries.
Practical tips¶
- Keep section requests narrow for better focus.
- Start with one section, then add another only if needed.
- Use
searchfirst when symbol ambiguity is possible.