kegg-analysis

KEGG Bioinformatics Analysis

Guide the user through structured KEGG database queries using the kegg MCP server tools.

Available Tools

You have access to 34 KEGG tools via the kegg MCP server. Key categories:

Category	Tools	Use for
Pathways	search_pathways, get_pathway_info, get_pathway_genes, get_pathway_compounds, get_pathway_reactions	Finding and exploring metabolic/signaling pathways
Genes	search_genes, get_gene_info, get_gene_orthologs	Gene function, cross-species orthologs
Compounds	search_compounds, get_compound_info, get_compound_reactions	Metabolites, substrates, products
Reactions	search_reactions, get_reaction_info	Biochemical transformations
Enzymes	search_enzymes, get_enzyme_info	EC numbers, catalytic activity
Diseases	search_diseases, get_disease_info	Disease-gene-drug associations
Drugs	search_drugs, get_drug_info, get_drug_interactions	Pharmacology, DDI
Orthology	search_ko_entries, get_ko_info	Functional orthologs (KO)
Cross-DB	batch_entry_lookup, convert_identifiers, find_related_entries	Bulk queries, ID mapping (UniProt, NCBI, ChEBI)
Visualization	render_pathway_ascii	ASCII pathway diagrams

Workflow Patterns

Pattern 1: Gene List → Pathway Enrichment

When the user provides a gene list:

1. Identify the organism code (e.g., hsa for human, eco for E. coli, sce for yeast)
2. For each gene, use search_genes with the organism to get KEGG gene IDs
3. For each gene ID, use get_gene_info with detail_level="full" to get pathway associations
4. Tally pathway frequencies across the gene list
5. For enriched pathways (appearing 2+ times), use get_pathway_info to describe their function
6. Summarize: which pathways are over-represented, what biological processes they reflect

Pattern 2: Pathway Deep Dive

When the user asks about a specific pathway:

1. Use search_pathways to find the pathway ID if not provided
2. Use get_pathway_info with detail_level="full" for overview
3. Use get_pathway_genes, get_pathway_compounds, get_pathway_reactions for components
4. Use render_pathway_ascii to visualize topology
5. Highlight key enzymes, rate-limiting steps, and regulatory points

Pattern 3: Cross-Species Comparison

When comparing metabolism across organisms:

1. Get the reference pathway (e.g., map00010 for glycolysis)
2. For each organism code, use get_pathway_info with the organism-specific ID (e.g., hsa00010, eco00010)
3. Use get_pathway_genes for each organism to list genes
4. Compare: which genes are conserved vs. species-specific
5. Use get_gene_orthologs for genes of interest to find KO assignments

Pattern 4: Compound/Drug Investigation

When investigating a compound or drug:

1. Use search_compounds or search_drugs to find the entry
2. Use get_compound_info or get_drug_info with detail_level="full"
3. Use get_compound_reactions to find biochemical context
4. Use find_related_entries to cross-link to pathways, enzymes, diseases
5. For drugs: use get_drug_interactions to screen DDI

Pattern 5: ID Mapping

When the user has identifiers from other databases:

1. Use convert_identifiers to map between KEGG and external DBs (UniProt, NCBI GeneID, ChEBI, PubChem)
2. Use batch_entry_lookup for bulk retrieval (up to 50 entries)
3. Use find_related_entries to traverse links between KEGG databases

KEGG Identifier Formats

Database	Format	Example
Pathway	map{5digits} or {org}{5digits}	map00010, hsa00010
Gene	{org}:{id}	hsa:1956, eco:b0002
Compound	C{5digits}	C00001 (water), C00002 (ATP)
Reaction	R{5digits}	R00756
Enzyme	ec:{EC number}	ec:1.1.1.1
KO	K{5digits}	K00001
Drug	D{5digits}	D00001
Disease	H{5digits}	H00004
Module	M{5digits}	M00001
Glycan	G{5digits}	G00001

Common Organism Codes

Code	Organism
hsa	Homo sapiens (human)
mmu	Mus musculus (mouse)
eco	Escherichia coli K-12
sce	Saccharomyces cerevisiae (yeast)
ath	Arabidopsis thaliana
dme	Drosophila melanogaster

Use list_organisms to find any organism code.

Tips

• Start with detail_level="summary" (default) to keep responses concise; switch to "full" when the user needs complete details
• Use batch_entry_lookup instead of individual calls when fetching 3+ entries
• render_pathway_ascii in grid mode shows spatial relationships; chain mode shows linear flow
• KEGG cross-references are bidirectional — use find_related_entries to traverse in either direction
• When the user provides gene symbols (e.g., EGFR, TP53), search with the organism code to resolve to KEGG IDs first