Tooluniverse Gene Enrichment

Runs GO, KEGG, and Reactome pathway enrichment on gene lists via gseapy or clusterProfiler. Built to solve the two big headaches in enrichment analysis: the canonical notebooks always use pre-computed results that you can't reproduce exactly (because the DEG filter or library snapshot was slightly different), and top hits constantly tie at the same adjusted p-value with no obvious winner. Ships three CLI scripts that handle tie-breaking deterministically, track candidate term ranks through simplify operations, and screen multiple conditions at once. The compute-first philosophy is solid: it immediately checks for executed notebooks and pre-computed CSVs before re-running anything, which saves you from chasing phantom discrepancies when the published answer is already sitting in the data folder.

skills.sh installs: 390
GitHub stars: 1.8K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add mims-harvard/tooluniverse --skill tooluniverse-gene-enrichment --agent claude-code

Installs into .claude/skills of the current project.

Files

COMPUTE, DON'T DESCRIBE

When analysis requires computation (statistics, data processing, scoring, enrichment), write and run Python code via Bash. Don't describe what you would do — execute it and report actual results. Use ToolUniverse tools to retrieve data, then Python (pandas, scipy, statsmodels, matplotlib) to analyze it.

Gene Enrichment and Pathway Analysis

RULE ZERO — Check for pre-computed results FIRST

Before following any instruction below, scan the data folder for:

*_executed.ipynb → read with tu run read_executed_notebook '{"data_folder":"<path>","search":"<keyword>"}' and cite its cell outputs as the authoritative answer
Pre-computed enrichment files (CSV/TSV named *enrich*, *go*, *kegg*, *reactome*, *ego*, *_simplified.csv) → read directly
Canonical analysis scripts (analysis.R, run_*.py, find_*.R, *.Rmd) → execute as-is and read the output

Only follow this skill's re-analysis recipe below if none of the above exist. Re-running enrichment from raw DEG lists produces different numbers than the published answer due to subtle filter differences upstream, and is much slower.