Skip to content

tooluniverse-gene-enrichment

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History
 
 

tooluniverse-gene-enrichment

README.md

tooluniverse-gene-enrichment

Comprehensive gene enrichment and pathway analysis skill for the ToolUniverse ecosystem.

Overview

Perform publication-ready gene enrichment analysis using Over-Representation Analysis (ORA) and Gene Set Enrichment Analysis (GSEA). Integrates local computation via gseapy with ToolUniverse APIs (PANTHER, STRING, Reactome) for cross-validated results across 225+ gene set libraries.

Capabilities

Capability Method Tool
GO Enrichment (BP, MF, CC) ORA / GSEA gseapy, PANTHER, STRING
KEGG Pathway Enrichment ORA / GSEA gseapy, STRING
Reactome Pathway Enrichment ORA gseapy, ReactomeAnalysis API
WikiPathways Enrichment ORA gseapy
MSigDB Hallmark Enrichment ORA / GSEA gseapy
Multiple Testing Correction BH, Bonferroni, BY, Holm statsmodels
Gene ID Conversion Symbol/Ensembl/Entrez/UniProt MyGene, STRING
Cross-Validation Multi-source consensus gseapy + PANTHER + STRING
Network Context PPI enrichment STRING
Multi-Organism human, mouse, rat, fly, worm, yeast, zebrafish All tools

Quick Start

import gseapy

# Standard GO enrichment (ORA)
result = gseapy.enrichr(
    gene_list=["TP53", "BRCA1", "EGFR", "MYC", "AKT1", "PTEN"],
    gene_sets='GO_Biological_Process_2021',
    organism='human',
    outdir=None,
    no_plot=True,
)
sig = result.results[result.results['Adjusted P-value'] < 0.05]
print(sig[['Term', 'Adjusted P-value', 'Overlap']].head(5))

Enrichment Databases

  • 225+ Enrichr libraries via gseapy (GO, KEGG, Reactome, WikiPathways, MSigDB, DisGeNET, CellMarker, GTEx, ChEA, ENCODE, and more)
  • PANTHER over-representation analysis (curated GO, PANTHER Pathways, Reactome)
  • STRING functional enrichment (GO, KEGG, Reactome, WikiPathways, COMPARTMENTS, DISEASES, HPO)
  • Reactome Analysis Service pathway overrepresentation

Supported Organisms

Human (9606), Mouse (10090), Rat (10116), Drosophila (7227), C. elegans (6239), Yeast (4932), Zebrafish (7955)

Files

File Description
SKILL.md Complete skill specification with 8-phase workflow, tool reference, response formats
QUICK_START.md 10 use case examples with copy-paste code
test_gene_enrichment.py 50-test comprehensive test suite (100% pass rate)
README.md This file

Test Results

Total: 50 | Passed: 50 | Failed: 0 | Time: 63.4s
Pass Rate: 50/50 (100.0%)

Test Coverage

Phase Tests Description
gseapy ORA 9 GO BP/MF/CC, KEGG, Reactome, Hallmark, WikiPathways, multi-library, structure
GSEA 3 Preranked GO/KEGG, NES direction
PANTHER 4 GO BP/MF/CC, PANTHER Pathways
STRING 3 Functional enrichment, KEGG, PPI enrichment
Reactome API 2 Pathway enrichment, cross-validation
ID Conversion 3 MyGene batch, STRING map, enrichment after conversion
Multiple Testing 3 BH, Bonferroni, method comparison
Cross-Validation 2 GO BP consensus, KEGG consensus
Gene Lists 4 Immune, signaling, small (3), large (43)
Term Lookup 2 Specific GO term, specific KEGG pathway
GO Details 2 Term details, gene annotations
Pathway Details 3 Reactome, KEGG, WikiPathways
Organisms 2 Mouse gseapy, PANTHER mouse
Library Discovery 2 List libraries, version verification
BixBench 4 Adjusted p-val, most enriched, GSEA metabolic, enrichGO equivalent
Integration 2 Full pipeline, comparative enrichment

BixBench Question Support

This skill is designed to answer BixBench-style enrichment questions:

  1. "What is the adjusted p-val for [GO term] from enrichGO enrichment analysis?"

    • Use gseapy.enrichr() with GO library, search results for specific term

  2. "Using gseapy and [library], which pathway is most significantly enriched?"

    • Use gseapy.enrichr() with specified library, sort by Adjusted P-value

  3. "In a KEGG pathway enrichment analysis, what is the p-value for the most enriched pathway?"

    • Use gseapy.enrichr() with KEGG_2021_Human, report top result

  4. "What GO Biological Processes are enriched (BH adjusted p < 0.05)?"

    • Use gseapy.enrichr(), filter by Adjusted P-value < 0.05

  5. "Compare enrichment using Bonferroni vs BH correction"

    • Use statsmodels.stats.multitest.multipletests() with both methods

Dependencies

  • Required: gseapy, pandas, scipy, statsmodels
  • ToolUniverse: PANTHER, STRING, Reactome, MyGene, GO, QuickGO, WikiPathways, KEGG tools
  • Optional: numpy (for GSEA ranked lists)

Installation

pip install gseapy pandas scipy statsmodels