Imported from davepoon/buildwithclaude (plugins/kegg-mcp-server/skills/kegg-analysis/SKILL.md). Install upstream with npx skills add davepoon/buildwithclaude --skill kegg-analysis. Copyright stays with the author.

KEGG Bioinformatics Analysis

This skill orchestrates multi-step biological analyses using the KEGG MCP server tools. It transforms raw gene lists, drug names, or pathway IDs into structured biological insights.

When to Use This Skill

• Performing pathway enrichment analysis on a gene list
• Investigating a drug's mechanism of action, targets, and interactions
• Comparing metabolic pathways across species
• Tracing compound-reaction networks
• Mapping genes to functional modules and ortholog groups

What This Skill Does

1. Identifies the analysis type from the user's input (enrichment, drug, comparison, network)
2. Resolves identifiers — maps gene symbols, drug names, or pathway IDs to KEGG entries
3. Retrieves cross-linked data — follows relationships across KEGG databases
4. Aggregates and ranks results — counts pathway hits, scores conservation, groups by function
5. Synthesizes biological context — explains significance, not just IDs

How to Use

Pathway Enrichment

Analyze these genes for pathway enrichment in human: BRCA1, TP53, EGFR, KRAS, PIK3CA

Workflow:

1. search_genes for each gene in the target organism (e.g., hsa)
2. get_gene_info to confirm identity and get KEGG gene IDs
3. find_related_entries to get pathway associations per gene
4. Aggregate: count how many input genes map to each pathway
5. get_pathway_info for top pathways
6. render_pathway_ascii for visual context
7. Report ranked pathways with p-value proxy (gene count / pathway size)

Drug Target Investigation

Investigate metformin: targets, pathways, and interactions

Workflow:

1. search_drugs to find the KEGG drug entry
2. get_drug_info for targets, classification, and metabolism
3. search_genes for each target gene
4. find_related_entries to get target pathways
5. get_drug_interactions for DDI screening
6. Synthesize mechanism-of-action summary

Cross-Species Comparison

Compare glycolysis (map00010) between human, E. coli, and yeast

Workflow:

1. get_pathway_info for organism-specific variants (hsa00010, eco00010, sce00010)
2. get_pathway_genes for each organism
3. get_gene_orthologs to identify conserved vs. species-specific enzymes
4. get_pathway_compounds to compare metabolite pools
5. render_pathway_ascii for each organism
6. Report conservation matrix and unique adaptations

Example

User: "What pathways are enriched in this gene set: SOD1, SOD2, CAT, GPX1, PRDX1?"

Output:

Pathway Enrichment Results (Homo sapiens)

Top Pathways:
1. hsa04146 Peroxisome (4/5 genes) — organelle for fatty acid oxidation and ROS detox
2. hsa04216 Ferroptosis (3/5 genes) — iron-dependent cell death regulated by GPX
3. hsa05022 Pathways of neurodegeneration (3/5 genes) — oxidative damage in ALS, AD, PD
4. hsa00480 Glutathione metabolism (2/5 genes) — GSH-dependent antioxidant system

Biological Context:
All 5 genes encode antioxidant enzymes. The enrichment in Peroxisome
and Ferroptosis pathways reflects their central role in reactive oxygen
species (ROS) detoxification. The neurodegeneration hit is consistent
with oxidative stress as a driver of SOD1-linked ALS.

Tips

• Provide organism context (human, mouse, E. coli) for faster resolution
• Use standard gene symbols — KEGG resolves HGNC symbols for human
• For large gene lists (>20), batch with batch_entry_lookup (max 50 per call)
• Cross-reference with convert_identifiers to bridge UniProt, NCBI Gene, or PDB IDs
• Use find_related_entries to discover unexpected connections between databases