LABSCO SUMMARY

Enrichment analysis with the two things that usually go wrong — library choice and background — handled deliberately.

Only significant terms come back, so nothing has to filter a table of thousands after the fact. The background option deserves a second read: Enrichr tests against the whole genome by default, and when your genes were drawn from a restricted universe — the genes expressed in one tissue, or a targeted panel — that default overstates significance, sometimes by orders of magnitude. Library choice is the other trap, and it is answered from Enrichr's live catalogue rather than a list frozen at release time.

What it is

A server over the Enrichr API. Hand it a gene list and it queries several gene-set libraries in parallel, returning only the terms that pass an adjusted p-value threshold; a companion tool picks which libraries to query for the question you are actually asking.

What you get

enrichr_analysis takes gene symbols — two at minimum — queries the chosen libraries in parallel, and returns only statistically significant terms with their p-values, odds ratio, combined score and overlapping genes
suggest_libraries searches Enrichr's live catalogue for libraries relevant to a research question, filtered by category and ranked, preferring the newer vintage when two are equally relevant
Because that catalogue is fetched at runtime rather than baked in, new libraries appear on their own and retired ones are never suggested
Background takes your own background gene set — at least 20 genes — instead of the whole genome, and the result reports backgroundCorrected per library so you can see it applied
Format switches the output between detailed, compact and minimal to control how much of the result reaches the context, and maxTerms caps terms per library
outputFile writes the complete results to a TSV file while the conversation keeps only the trimmed version
Every call returns formatted text alongside structured JSON with the full result data

Requirements

Published as enrichr-mcp-server and launched over stdio with npx. No account and no key — Enrichr's public API is all it reaches. A .mcpb bundle is published for Claude Desktop, and the one-click installs for Cursor and VS Code pass --compact and --max-terms 100 by default.

Setup effort

One command — npx -y enrichr-mcp-server

See installation options

USE CASES

01

Making sense of a differential-expression hit list

Ask which libraries suit the biology in question, then run the gene list against them in one parallel pass.

suggest_librariesenrichr_analysis

02

Testing against the universe your assay actually covered

Pass the expressed or panel genes as the background so significance is measured against what could have been detected.

enrichr_analysis

03

Picking libraries for a specific topic

A ranked, category-filtered list from the live catalogue, so nothing retired is recommended and the newer vintage wins ties.

suggest_libraries

04

Keeping a large result out of the conversation

Return the compact or minimal view to the model and write the complete table to a TSV file at the same time.

enrichr_analysis

2 tools

Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.

`enrichr_analysis`

Perform gene set enrichment analysis using Enrichr across multiple gene set libraries. Returns only statistically significant terms (adjusted p < 0.05).

Configured default libraries:

GO_Biological_Process_2026: Current Gene Ontology biological process terms — biological objectives and programs accomplished by gene products. The default choice for GO biological process enrichment.
KEGG_2026: Current KEGG metabolic and signaling pathways — the up-to-date KEGG release; prefer over older KEGG vintages.
Reactome_Pathways_2024: Current Reactome release — curated, peer-reviewed pathways for signaling, metabolism, gene expression, and disease.
MSigDB_Hallmark_2020: Hallmark gene sets representing well-defined biological states and processes from MSigDB.
ChEA_2022: ChIP-seq experiments from GEO, ENCODE, and publications identifying transcription factor-gene interactions from human and mouse.
GWAS_Catalog_2025: Current genome-wide association study results from the NHGRI-EBI GWAS Catalog, linking genes to human traits and diseases.
Human_Phenotype_Ontology: Standardized vocabulary of phenotypic abnormalities associated with human diseases.
PPI_Hub_Proteins: Highly connected hub proteins from protein-protein interaction networks.
DGIdb_Drug_Targets_2024: Drug-gene interactions from the Drug Gene Interaction Database — druggable targets and their compounds.
CellMarker_2024: Manually curated cell type markers from CellMarker database for human and mouse.

Any library in Enrichr's live catalog is accepted. Enrichr adds and retires libraries continuously, so do not rely on a memorized list: call suggest_libraries, or read the enrichr://libraries resource, to discover the libraries that currently exist.

Pass 'background' to test against a custom background gene set (for example, only the genes expressed in your assay) instead of Enrichr's whole-genome default. This is the statistically correct choice whenever the gene list was drawn from a restricted universe, and it matters: the whole-genome default can overstate significance by many orders of magnitude.

Each library's result reports 'backgroundCorrected'. If Enrichr's background service is unavailable the result falls back to uncorrected whole-genome p-values, flagged with a warning — treat those numbers as inflated and re-run rather than reporting them as background-corrected.

BEFORE YOU INSTALL

What to know first

Tools you get

2 tools when we connected.

Requires

Node.js 18+

Where it runs

Both documented clients run the server on your machine.

CONTEXT COST

Light Standing cost

Rated by the size of the 2 tool definitions that ride along on every turn, not by how many there are

INSTALLATION

Connect your client

Server setup

TERMINAL

npx -y enrichr-mcp-server

MCP Clients

Cursor

CURSOR · JSON

{
  "mcpServers": {
    "enrichr-server": {
      "command": "npx",
      "args": [
        "-y",
        "enrichr-mcp-server"
      ]
    }
  }
}

Claude Code

CLAUDE CODE · TERMINAL

claude mcp add enrichr-mcp-server -- npx -y enrichr-mcp-server

Enrichr MCP Server

Enrichment analysis with the two things that usually go wrong — library choice and background — handled deliberately.

What it is

What you get

Requirements

Setup effort

Making sense of a differential-expression hit list

Testing against the universe your assay actually covered

Picking libraries for a specific topic

Keeping a large result out of the conversation

2 tools

`enrichr_analysis`

`suggest_libraries`

What to know first

Tools you get

Requires

Where it runs

Light Standing cost

Connect your client

Server setup

MCP Clients

Cursor

Claude Code

Enrichr MCP Server

Enrichment analysis with the two things that usually go wrong — library choice and background — handled deliberately.

What it is

What you get

Requirements

Setup effort

Making sense of a differential-expression hit list

Testing against the universe your assay actually covered

Picking libraries for a specific topic

Keeping a large result out of the conversation

2 tools

enrichr_analysis

suggest_libraries

What to know first

Tools you get

Requires

Where it runs

Light Standing cost

Connect your client

Server setup

MCP Clients

Cursor

Claude Code

Same category

SCMCP

Unofficial Gene Ontology MCP Server

Unofficial STRING MCP Server

Bioinformatics MCP Server

`enrichr_analysis`

`suggest_libraries`