hgnc-link
Grounds gene-nomenclature work in the HUGO Gene Nomenclature Committee (HGNC) dataset, enabling resolution of gene symbols and IDs to canonical HGNC identifiers, plus cross-references and batch operations.
README
hgnc-link
An MCP (Model Context Protocol) + HTTP server that grounds gene-nomenclature work in the HUGO Gene Nomenclature Committee (HGNC) dataset from genenames.org.
It is a sibling of uniprot-link, gnomad-link, gencc-link, clingen-link,
etc., and shares their architecture: a FastAPI + FastMCP unified server, a typed
error envelope, {tool, arguments} chaining via _meta.next_commands, and
capabilities-as-resource discovery.
Why
Every downstream genetics tool needs the same thing from HGNC: turn any gene
symbol — current, outdated (previous), or alias — and any HGNC ID form into the
canonical {hgnc_id, approved_symbol}, then pull cross-references. hgnc-link
makes that a single fast call, with the match provenance and any ambiguity made
explicit instead of silently collapsed.
Speed: local index, refreshed by cron
For speed and reliability the server is backed by a local SQLite index built
from HGNC's bulk downloads (hgnc_complete_set.json + withdrawn.txt) and
refreshed by a cron-invoked CLI — no per-request REST round-trips. The live
rest.genenames.org API is used only as an optional fallback before the first
build completes.
# Build the index once (downloads ~33 MB, builds in seconds):
uv run hgnc-link-data build
# Cron entry point — conditional refresh (304-cheap; rebuilds only on change):
uv run hgnc-link-data refresh
# Inspect the loaded release:
uv run hgnc-link-data status
See docs/deployment.md for a crontab line and a systemd
timer.
Tools
| Tool | Purpose |
|---|---|
get_server_capabilities |
Discovery surface (tools, signatures, workflows, vocab). |
get_hgnc_diagnostics |
Loaded release, counts, freshness, data-source status. |
resolve_symbol |
Any symbol/ID → {hgnc_id, approved_symbol, match_type} + candidates. |
resolve_symbols_batch |
Resolve many symbols/IDs at once (never fails on a miss). |
get_gene |
Full HGNC record (alias/previous aware). |
search_genes |
FTS over symbol/name/alias/previous symbols. |
get_gene_cross_references |
Gene → NCBI/Ensembl/UniProt/RefSeq/MANE/OMIM/… |
lookup_by_xref |
External ID → HGNC gene (reverse mapping). |
get_gene_group |
Browse a gene family by group ID or name. |
Every response carries _meta.next_commands (a ready-to-call {tool, arguments}
list) on success and error. Response verbosity is controlled by
response_mode ∈ {minimal, compact, standard, full} (default compact).
Quick start
make install # uv sync --group dev
make data # build the local HGNC index
make dev # unified REST + MCP server on http://127.0.0.1:8000/mcp
make mcp-serve # stdio MCP server (Claude Desktop)
make test # unit tests
Register with Claude Code (HTTP):
claude mcp add --transport http hgnc-link --scope user http://127.0.0.1:8000/mcp
Or stdio (Claude Desktop) — see claude-desktop-config.json.
Data & license
HGNC data is released with no usage restrictions (effectively CC0). Attribution is requested: Seal RL, et al. Genenames.org: the HGNC resources in 2023. Nucleic Acids Res. RRID:SCR_002827.
Research use only; not for clinical decision support.
Development
make check # format + lint
make typecheck # mypy --strict
make ci-local # format-check + lint + lint-loc + typecheck + tests
make test-integration # live HGNC download + REST asserts (opt-in)
Architecture details: docs/architecture.md.
Design spec: docs/superpowers/specs/2026-06-12-hgnc-link-design.md.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。