🛤️ KEGG
HOME

🧠 Pathway Analysis Fundamentals

🎯 What is Pathway Analysis?

Pathway analysis is the process of interpreting high-throughput molecular data (genes, proteins, metabolites) in the context of biological pathways. It identifies which pathways are significantly affected in a condition, revealing the functional meaning behind gene lists. KEGG (Kyoto Encyclopedia of Genes and Genomes) is one of the most widely used pathway databases, providing maps for metabolic, signaling, and disease pathways across thousands of organisms.

Why Pathway Analysis Matters

Interpretation: Biological meaning. Enrichment: Over-represented pathways. Mechanism: How genes work together. Comparison: Across conditions. Drug discovery: Targets. Systems: Holistic view.

📌 Interpretation📌 Enrichment📌 Mechanism📌 Systems
# KEGG & Pathway Analysis — the big picture # Interpreting omics data in pathway context. print(""" === KEGG & Pathway Analysis === What it is: - Interpret omics data - Biological pathways - Over-representation - Functional meaning Why it matters: - Biological interpretation - Pathway enrichment - Mechanism of action - Cross-condition comparison - Drug discovery - Systems biology KEGG: - Kyoto Encyclopedia of Genes and Genomes - Pathways, genes, compounds, diseases - Cross-species - KO numbers Other databases: - Reactome - WikiPathways - BioCyc - PANTHER - Pathway Commons Applications: - Transcriptomics - Proteomics - Metabolomics - Multi-omics - Disease analysis - Drug discovery """) print("=== Historical Context ===") print(""" 1995: KEGG founded (Kanehisa) 2000: KEGG PATHWAY database 2005: KEGG BRITE, KO 2010: Pathway enrichment tools 2015: GSEA, clusterProfiler 2020: Multi-omics, single-cell 2025: AI, knowledge graphs, integration """) print("=== Why It's Hard ===") print(""" 1. COVERAGE - Not all genes in pathways - Incomplete annotation 2. CROSS-SPECIES - Different gene IDs - Orthology mapping 3. REDUNDANCY - Overlapping pathways - Hierarchical 4. STATISTICS - Multiple testing - Background choice 5. VISUALIZATION - Complex maps - Many genes """)

🎯 Why KEGG Matters

🎯 The Value of KEGG

KEGG provides a comprehensive, cross-species, manually curated resource for pathway analysis. It's the most widely used pathway database in bioinformatics.

Key Benefits

Comprehensive: Many pathways. Cross-species: Orthology. Curated: Manual. Visual: Maps. Integrated: Genes, compounds, diseases. Free: Academic.

🎯 Comprehensive🎯 Cross-species🎯 Curated🎯 Visual
# Why KEGG Matters print("=== 1. Comprehensive ===") print(""" Comprehensive: - > 500 pathways - Metabolic, signaling, disease - > 8,000 organisms - > 30 million genes Example: Glycolysis: hsa00010 MAPK signaling: hsa04010 p53 signaling: hsa04115 Cancer: hsa05200 """) print("\n=== 2. Cross-Species ===") print(""" Cross-species: - KO numbers - Orthology - Compare species Example: Human gene TP53 → KO K04451 → Mouse gene Trp53 → Same KO → Same pathway """) print("\n=== 3. Curated ===") print(""" Curated: - Manual annotation - Expert review - Quality Example: KEGG pathway maps → Manually drawn → Updated regularly """) print("\n=== 4. Visual ===") print(""" Visual: - Pathway maps - Diagrams - Clear Example: Glycolysis map → Enzymes → Metabolites → Reactions """) print("\n=== 5. Integrated ===") print(""" Integrated: - Genes - Compounds - Reactions - Diseases - Drugs Example: Pathway → Genes → Diseases → Drug targets """) print("\n=== 6. Free ===") print(""" Free: - Academic use - Web access - API - FTP Example: https://www.kegg.jp/ https://rest.kegg.jp/ """) print("\n=== 7. Enrichment ===") print(""" Enrichment: - Gene list → pathways - Over-representation - Statistical Example: DE genes → Enriched pathways → Biological meaning """) print("\n=== 8. Drug Discovery ===") print(""" Drug: - Target identification - Pathway analysis - Drug repurposing Example: Cancer pathways → Drug targets → Therapeutics """) print("\n=== 9. Disease Analysis ===") print(""" Disease: - Disease pathways - Mechanisms - Biomarkers Example: Alzheimer's → Pathways → Mechanisms """) print("\n=== 10. Systems Biology ===") print(""" Systems: - Networks - Pathways - Integration Example: Multi-omics → Pathway analysis → Systems view """)

🗺️ KEGG Database

🎯 KEGG Resources

KEGG is a collection of databases: PATHWAY, GENES, KO, BRITE, COMPOUND, DRUG, DISEASE, and more. Each serves a different purpose.

KEGG Databases

PATHWAY: Maps. GENES: Genomes. KO: Orthology. BRITE: Hierarchy. COMPOUND: Chemicals. DISEASE: Diseases.

🗺️ PATHWAY🗺️ GENES🗺️ KO🗺️ BRITE
# KEGG Database print("=== KEGG PATHWAY ===") print(""" PATHWAY: - Pathway maps - Metabolic, signaling, disease - > 500 pathways - Manually drawn Example: hsa00010 — Glycolysis hsa04010 — MAPK signaling hsa04115 — p53 signaling hsa05200 — Cancer Format: - KGML (XML) - PNG (image) - HTML (web) """) print("\n=== KEGG GENES ===") print(""" GENES: - Complete genomes - > 8,000 organisms - Gene catalogs Example: hsa:7157 — Human TP53 mmu:22059 — Mouse Trp53 dme:TP53 — Fly p53 Format: - FASTA - Flat file """) print("\n=== KEGG KO ===") print(""" KO (KEGG Orthology): - Orthologous genes - Functional groups - Cross-species Example: K04451 — TP53 K00001 — Alcohol dehydrogenase Format: - K number - Functional annotation """) print("\n=== KEGG BRITE ===") print(""" BRITE: - Hierarchical classifications - Functional hierarchies - Protein families Example: ko00001 — KEGG Orthology ko03000 — Transcription factors ko04131 — Membrane trafficking Format: - Hierarchical - Text """) print("\n=== KEGG COMPOUND ===") print(""" COMPOUND: - Chemical compounds - Metabolites - Drugs Example: C00031 — D-Glucose C00002 — ATP Format: - C number - Chemical structure """) print("\n=== KEGG DRUG ===") print(""" DRUG: - Approved drugs - Drug targets - Drug metabolism Example: D00001 — Water D00123 — Aspirin Format: - D number - Drug information """) print("\n=== KEGG DISEASE ===") print(""" DISEASE: - Disease entries - Disease genes - Pathways Example: H00001 — Acute myeloid leukemia H00002 — Bladder cancer Format: - H number - Disease information """) print("\n=== KEGG Modules ===") print(""" MODULE: - Functional units - Sub-pathways - Building blocks Example: M00001 — Glycolysis M00002 — Gluconeogenesis Format: - M number - Functional module """) print("\n=== Database Statistics ===") print(f"{'Database':<20} {'Entries':<15} {'Description':<30}") print("-" * 65) print(f"{'PATHWAY':<20} {'~500':<15} {'Pathway maps':<30}") print(f"{'GENES':<20} {'~30M':<15} {'Genes':<30}") print(f"{'KO':<20} {'~25,000':<15} {'Orthologs':<30}") print(f"{'BRITE':<20} {'~1,000':<15} {'Hierarchies':<30}") print(f"{'COMPOUND':<20} {'~20,000':<15} {'Compounds':<30}") print(f"{'DRUG':<20} {'~12,000':<15} {'Drugs':<30}") print(f"{'DISEASE':<20} {'~2,500':<15} {'Diseases':<30}") print("\n=== Access ===") print(""" Web: - https://www.kegg.jp/ API: - https://rest.kegg.jp/ FTP: - ftp://ftp.genome.jp/pub/kegg/ """)

🔢 KO & BRITE

🎯 KEGG Orthology

KO (KEGG Orthology) is a classification system that groups orthologous genes from different species into functional units. Each KO has a unique identifier (K number) and links genes to pathways.

KO & BRITE

KO: Ortholog group. K number: ID. BRITE: Hierarchy. Mapping: Gene → KO. Cross-species: Compare. Pathways: Link.

🔢 KO🔢 K number🔢 BRITE🔢 Cross-species
# KO & BRITE print("=== KO (KEGG Orthology) ===") print(""" KO: - Orthologous genes - Functional group - Cross-species Example: K04451 — TP53 K00001 — Alcohol dehydrogenase Format: - K number - Functional annotation - Gene links """) print("\n=== KO Mapping ===") print(""" Mapping: - Gene → KO - Species-specific - Orthology Example: Human TP53 (hsa:7157) → KO K04451 Mouse Trp53 (mmu:22059) → KO K04451 Fly p53 (dme:TP53) → KO K04451 """) print("\n=== BRITE ===") print(""" BRITE: - Hierarchical classifications - Functional hierarchies - Protein families Example: ko00001 — KEGG Orthology ko03000 — Transcription factors ko04131 — Membrane trafficking ko02000 — Transporters Format: - Hierarchical - Text - Gene lists """) print("\n=== KO Hierarchy ===") print(""" Hierarchy: - ko00001 (KEGG Orthology) ├── Metabolism │ ├── Carbohydrate metabolism │ ├── Energy metabolism │ └── ... ├── Genetic Information Processing ├── Environmental Information Processing ├── Cellular Processes ├── Organismal Systems └── Human Diseases """) print("\n=== KO Statistics ===") print(""" Statistics: - KO entries: ~25,000 - Genes: ~30 million - Species: ~8,000 - Pathways: ~500 """) print("\n=== KO Retrieval ===") print(""" Retrieval: - KEGG API - KEGG website - KEGGREST (R) - Requests (Python) Example: https://rest.kegg.jp/get/K04451 https://rest.kegg.jp/link/hsa/K04451 """) print("\n=== Python KO Retrieval ===") print(""" import requests def get_ko(ko_id): url = f"https://rest.kegg.jp/get/{ko_id}" response = requests.get(url) if response.status_code == 200: return response.text return None def get_ko_genes(ko_id, organism="hsa"): url = f"https://rest.kegg.jp/link/{organism}/{ko_id}" response = requests.get(url) if response.status_code == 200: return response.text return None # Example ko = get_ko("K04451") print(ko) genes = get_ko_genes("K04451", "hsa") print(genes) """) print("\n=== KO to Pathway ===") print(""" KO → Pathway: - Link KO to pathways - Cross-reference Example: K04451 (TP53) → hsa04115 (p53 signaling) → hsa05200 (Cancer) → hsa04210 (Apoptosis) """) print("\n=== Use Cases ===") print(""" Use cases: - Cross-species comparison - Functional annotation - Pathway mapping - Orthology assignment - Metagenomics - Comparative genomics """)

🛤️ Pathway Types

🎯 Types of KEGG Pathways

KEGG pathways are organized into several categories: metabolism, genetic information processing, environmental information processing, cellular processes, organismal systems, and human diseases.

Pathway Categories

Metabolism: Biochemical. Genetic: DNA/RNA. Environmental: Signaling. Cellular: Processes. Organismal: Systems. Disease: Human.

🛤️ Metabolism🛤️ Genetic🛤️ Environmental🛤️ Disease
# Pathway Types print("=== 1. Metabolism ===") print(""" Metabolism: - Carbohydrate - Energy - Lipid - Nucleotide - Amino acid - Glycan - Cofactor - Secondary metabolites Examples: hsa00010 — Glycolysis hsa00020 — TCA cycle hsa00190 — Oxidative phosphorylation hsa00980 — Metabolism of xenobiotics """) print("\n=== 2. Genetic Information Processing ===") print(""" Genetic: - Transcription - Translation - Replication - Repair - Folding - Degradation Examples: hsa03010 — Ribosome hsa03040 — Spliceosome hsa03410 — Base excision repair hsa04141 — Protein processing """) print("\n=== 3. Environmental Information Processing ===") print(""" Environmental: - Signal transduction - Membrane transport - Signaling molecules Examples: hsa04010 — MAPK signaling hsa04151 — PI3K-Akt signaling hsa04310 — Wnt signaling hsa02010 — ABC transporters """) print("\n=== 4. Cellular Processes ===") print(""" Cellular: - Cell cycle - Cell death - Cell motility - Transport - Catabolism Examples: hsa04110 — Cell cycle hsa04210 — Apoptosis hsa04144 — Endocytosis hsa04140 — Autophagy """) print("\n=== 5. Organismal Systems ===") print(""" Organismal: - Immune - Endocrine - Circulatory - Digestive - Nervous - Sensory - Development - Aging Examples: hsa04610 — Complement cascade hsa04910 — Insulin signaling hsa04080 — Neuroactive ligand hsa04211 — Longevity """) print("\n=== 6. Human Diseases ===") print(""" Disease: - Cancer - Infectious - Immune - Neurodegenerative - Cardiovascular - Metabolic Examples: hsa05200 — Cancer hsa05130 — Pathogenic E. coli hsa05010 — Alzheimer's hsa04930 — Type II diabetes """) print("\n=== 7. Drug Development ===") print(""" Drug: - Drug targets - Drug metabolism - Drug action Examples: hsa00982 — Drug metabolism hsa07010 — Drug targets """) print("\n=== 8. Pathway IDs ===") print(""" Format: - Organism prefix (hsa, mmu, etc.) - 5-digit number - Category Examples: hsa00010 — Human glycolysis mmu00010 — Mouse glycolysis eco00010 — E. coli glycolysis """) print("\n=== 9. Pathway Size ===") print(f"{'Category':<30} {'Pathways':<15}") print("-" * 45) print(f"{'Metabolism':<30} {'~150':<15}") print(f"{'Genetic Information':<30} {'~30':<15}") print(f"{'Environmental Information':<30} {'~50':<15}") print(f"{'Cellular Processes':<30} {'~50':<15}") print(f"{'Organismal Systems':<30} {'~100':<15}") print(f"{'Human Diseases':<30} {'~100':<15}") print(f"{'Drug Development':<30} {'~10':<15}") print("\n=== 10. Use Cases ===") print(""" Use cases: - Metabolic analysis - Signaling analysis - Disease analysis - Drug discovery - Systems biology - Comparative genomics """)

📊 Pathway Enrichment

🎯 Finding Over-Represented Pathways

Pathway enrichment analysis identifies KEGG pathways that are over-represented in a set of genes compared to a background. It's the most common pathway analysis method.

Enrichment

Input: Gene list. Background: All genes. Test: Hypergeometric. Correction: FDR. Output: Enriched pathways. Interpret: Biology.

📊 Input📊 Background📊 Test📊 FDR
# Pathway Enrichment print("=== What is Enrichment? ===") print(""" Enrichment: - Gene list (e.g., DE genes) - Background (all genes) - Over-representation Example: 100 DE genes → 20 in "Cell cycle" → Expected: 5 → Enriched (p < 0.001) """) print("\n=== Statistical Test ===") print(""" Hypergeometric test: - Probability of observing k or more - Given N, K, n Parameters: - N: total genes - K: genes in pathway - n: gene list size - k: overlap Example: N = 20,000 K = 200 (cell cycle) n = 100 (DE genes) k = 20 (overlap) → p-value """) print("\n=== Multiple Testing ===") print(""" Multiple testing: - Many pathways tested - False positives - Correction needed Corrections: - Bonferroni: strict - Benjamini-Hochberg (FDR): common - Storey: less strict Example: 500 pathways → 25 false positives at p < 0.05 → FDR correction """) print("\n=== Background ===") print(""" Background: - All genes tested - Or all genes in genome - Affects results Example: RNA-seq: all expressed genes Microarray: all probes Genome: all genes Choice matters! """) print("\n=== Tools ===") print(""" 1. clusterProfiler (R) 2. KEGGREST (R) 3. Pathview (R) 4. DAVID (Web) 5. g:Profiler (Web) 6. Enrichr (Web) 7. WebGestalt (Web) 8. KEGG API (Python) 9. GSEA (Web/Desktop) """) print("\n=== R clusterProfiler ===") print(""" library(clusterProfiler) library(org.Hs.eg.db) # Enrichment ekegg <- enrichKEGG( gene = gene_list, organism = "hsa", pvalueCutoff = 0.05, qvalueCutoff = 0.05 ) # View head(ekegg) # Plot dotplot(ekegg) barplot(ekegg) """) print("\n=== Python KEGGREST ===") print(""" import requests from scipy.stats import hypergeom def kegg_enrichment(gene_list, background, organism="hsa"): \"\"\"Simple KEGG enrichment.\"\"\" # Get pathways for genes # ... pass """) print("\n=== Interpretation ===") print(""" Interpretation: - Enriched pathways - Biological meaning - Hypotheses Example: Cancer dataset → "Cell cycle" enriched → Proliferation → Hypothesis """) print("\n=== Common Pitfalls ===") print(""" Pitfalls: - Wrong background - Multiple testing - Redundant pathways - Small gene lists - Annotation bias Best practices: - Correct background - FDR correction - Filter redundant - Min gene list (10-20) - Consider bias """) print("\n=== Visualization ===") print(""" Visualization: - Bar plot - Dot plot - Pathway map - Network Tools: - clusterProfiler - pathview - KEGG mapper """)

📈 GSEA

🎯 Gene Set Enrichment Analysis

GSEA (Gene Set Enrichment Analysis) determines whether a defined set of genes (e.g., a KEGG pathway) shows statistically significant, concordant differences between two biological states. Unlike enrichment, GSEA uses all genes ranked by expression.

GSEA

Input: Ranked gene list. Sets: Pathways. Method: Running sum. Output: Enrichment score. Correction: FDR. Interpret: Coordinated changes.

📈 Ranked list📈 Gene sets📈 ES📈 FDR
# GSEA print("=== What is GSEA? ===") print(""" GSEA: - Gene Set Enrichment Analysis - Ranked gene list - Gene sets (pathways) - Coordinated changes Difference from enrichment: - Enrichment: cutoff list - GSEA: all genes ranked - GSEA: more sensitive """) print("\n=== Algorithm ===") print(""" Steps: 1. Rank genes by expression 2. For each gene set: - Calculate running sum - Enrichment score (ES) 3. Permutation test 4. Normalize (NES) 5. FDR correction Output: - ES: enrichment score - NES: normalized ES - p-value - FDR - Leading edge """) print("\n=== Enrichment Score ===") print(""" ES: - Maximum deviation from zero - Running sum - Positive: enriched at top - Negative: enriched at bottom Example: ES = 0.8 → Enriched at top → Upregulated """) print("\n=== Leading Edge ===") print(""" Leading edge: - Genes contributing to ES - Core enrichment - Most important Example: Pathway with 100 genes → 20 in leading edge → Key genes """) print("\n=== Tools ===") print(""" 1. GSEA (Broad) - Desktop - Web - Standard 2. clusterProfiler (R) - gseKEGG - gseGO 3. fgsea (R) - Fast - Flexible 4. GSEApy (Python) - Python - Flexible 5. WebGestalt - Web - Comprehensive """) print("\n=== R clusterProfiler GSEA ===") print(""" library(clusterProfiler) # GSEA gsea <- gseKEGG( geneList = ranked_genes, organism = "hsa", pvalueCutoff = 0.05 ) # View head(gsea) # Plot gseaplot(gsea, geneSetID = "hsa04110") """) print("\n=== Python GSEApy ===") print(""" import gseapy # GSEA gsea = gseapy.gsea( data=expression_data, gene_sets='KEGG_2021_Human', cls='phenotype.cls', permutation_num=1000, outdir='gsea_output' ) # View print(gsea.res2d) """) print("\n=== Interpretation ===") print(""" Interpretation: - Positive NES: upregulated - Negative NES: downregulated - Leading edge: key genes Example: Cell cycle: NES = 2.1 → Upregulated → Proliferation """) print("\n=== Advantages ===") print(""" Advantages: - No cutoff - More sensitive - Detects coordinated changes - Robust Disadvantages: - Slower - Complex - Requires ranking """) print("\n=== Use Cases ===") print(""" Use cases: - Transcriptomics - Proteomics - Drug response - Disease - Time course - Single-cell """)

🎨 Pathway Visualization

🎯 Seeing Pathways

Pathway visualization is crucial for interpreting results. Tools like pathview, KEGG Mapper, and clusterProfiler provide rich visualizations.

Visualization

Pathview: Map data. KEGG Mapper: Web. clusterProfiler: Plots. Cytoscape: Networks. iPath: Multi-omics. Python: Custom.

🎨 Pathview🎨 KEGG Mapper🎨 clusterProfiler🎨 Cytoscape
# Pathway Visualization print("=== 1. Pathview ===") print(""" Pathview: - R package - Maps data to KEGG pathway - Color by expression - Publication quality Example: library(pathview) pv.out <- pathview( gene.data = gene_data, pathway.id = "hsa04110", species = "hsa" ) """) print("\n=== 2. KEGG Mapper ===") print(""" KEGG Mapper: - Web tool - Maps genes to pathways - Color by data - Interactive URL: https://www.kegg.jp/kegg/mapper/ Use: - Upload gene list - Select pathway - Visualize """) print("\n=== 3. clusterProfiler ===") print(""" clusterProfiler: - R package - Dot plot - Bar plot - Enrichment map - Network Example: dotplot(ekegg) barplot(ekegg) cnetplot(ekegg) emapplot(ekegg) """) print("\n=== 4. Cytoscape ===") print(""" Cytoscape: - Network visualization - Pathway analysis - Enrichment map Example: - Import enrichment - EnrichmentMap - Clustering - Layout """) print("\n=== 5. iPath ===") print(""" iPath: - Web tool - Multi-omics - Pathway mapping - Interactive URL: https://pathways.embl.de/ """) print("\n=== 6. Python Visualization ===") print(""" import matplotlib.pyplot as plt import pandas as pd def plot_pathway_enrichment(results): \"\"\"Plot pathway enrichment.\"\"\" df = pd.DataFrame([{ "pathway": r["Description"], "pvalue": r["pvalue"], "count": r["Count"] } for r in results]) df = df.sort_values("pvalue").head(20) plt.figure(figsize=(10, 8)) plt.barh(df["pathway"], -np.log10(df["pvalue"]), color='purple') plt.xlabel("-log10(p-value)") plt.title("Pathway Enrichment") plt.tight_layout() plt.savefig("pathway_enrichment.png") plt.show() # Use # plot_pathway_enrichment(results) """) print("\n=== 7. Network Visualization ===") print(""" Network: - Nodes: Pathways - Edges: Overlap - Clusters: Modules Tools: - Cytoscape - NetworkX - igraph - clusterProfiler """) print("\n=== 8. Best Practices ===") print(""" 1. Sort by p-value 2. Show top 20 3. Color by p-value 4. Size by count 5. Remove redundancy 6. Label clearly 7. Export high-quality 8. Include legend """)

🛠️ Tools (clusterProfiler, pathview, KEGGREST)

🎯 Pathway Analysis Tools

Many tools exist for KEGG pathway analysis. Each has strengths and is suited to different use cases.

Tools

clusterProfiler: R. pathview: R. KEGGREST: R. GSEApy: Python. DAVID: Web. KEGG Mapper: Web.

🛠️ clusterProfiler🛠️ pathview🛠️ KEGGREST🛠️ GSEApy
# Tools print("=== clusterProfiler ===") print(""" clusterProfiler: - R package - Enrichment - GSEA - Visualization Features: - enrichKEGG - gseKEGG - dotplot - barplot - cnetplot - emapplot Install: BiocManager::install("clusterProfiler") """) print("\n=== pathview ===") print(""" pathview: - R package - Map data to pathway - Color by expression - Publication quality Features: - pathview() - KEGG pathway - Gene data - Compound data Install: BiocManager::install("pathview") """) print("\n=== KEGGREST ===") print(""" KEGGREST: - R package - KEGG API - Retrieve data - Programmatic Features: - keggGet - keggList - keggLink - keggFind Install: BiocManager::install("KEGGREST") """) print("\n=== GSEApy ===") print(""" GSEApy: - Python - GSEA - Enrichment - Flexible Features: - gsea - enrichr - prerank - Visualization Install: pip install gseapy """) print("\n=== DAVID ===") print(""" DAVID: - Web: david.ncifcrf.gov - Popular - Enrichment - Functional annotation Features: - KEGG enrichment - GO enrichment - Clustering - Visualization """) print("\n=== KEGG Mapper ===") print(""" KEGG Mapper: - Web: kegg.jp/kegg/mapper - Map genes - Color pathways - Interactive Features: - Search - Color - Export - Link """) print("\n=== Other Tools ===") print(""" 1. WebGestalt - Web - Comprehensive 2. g:Profiler - Web - Fast 3. Enrichr - Web - Simple 4. Reactome - Web - Pathways 5. WikiPathways - Web - Community 6. PANTHER - Web - Classification 7. iPathwayGuide - Web - Advanced """) print("\n=== Tool Comparison ===") print(f"{'Tool':<20} {'Type':<15} {'Enrichment':<15} {'GSEA':<15} {'Free':<10}") print("-" * 75) print(f"{'clusterProfiler':<20} {'R':<15} {'Yes':<15} {'Yes':<15} {'Yes':<10}") print(f"{'pathview':<20} {'R':<15} {'No':<15} {'No':<15} {'Yes':<10}") print(f"{'KEGGREST':<20} {'R':<15} {'No':<15} {'No':<15} {'Yes':<10}") print(f"{'GSEApy':<20} {'Python':<15} {'Yes':<15} {'Yes':<15} {'Yes':<10}") print(f"{'DAVID':<20} {'Web':<15} {'Yes':<15} {'No':<15} {'Yes':<10}") print(f"{'GSEA':<20} {'Desktop':<15} {'No':<15} {'Yes':<15} {'Yes':<10}") print("\n=== When to Use Which ===") print(""" USE clusterProfiler: ✓ R ✓ Enrichment ✓ GSEA ✓ Visualization USE pathview: ✓ Map data ✓ Publication ✓ KEGG pathway USE KEGGREST: ✓ Retrieve data ✓ Programmatic ✓ R USE GSEApy: ✓ Python ✓ GSEA ✓ Flexible USE DAVID: ✓ Quick ✓ No coding ✓ Popular USE KEGG Mapper: ✓ Web ✓ Interactive ✓ Visualization """)

🐍 Python Integration

🎯 Pathway Analysis with Python

Python is a powerful language for KEGG pathway analysis. Here's how to use it with KEGG API, GSEApy, and other libraries.

Python

Requests: KEGG API. GSEApy: GSEA. Pandas: Data. Matplotlib: Plot. NetworkX: Graphs. Automate: Pipeline.

🐍 Requests🐍 GSEApy🐍 Pandas🐍 Automate
# Python Integration print("=== 1. KEGG API with Requests ===") print(""" import requests def get_kegg_pathway(pathway_id): url = f"https://rest.kegg.jp/get/{pathway_id}" response = requests.get(url) if response.status_code == 200: return response.text return None # Example pathway = get_kegg_pathway("hsa04110") print(pathway[:500]) """) print("\n=== 2. Get Genes in Pathway ===") print(""" def get_pathway_genes(pathway_id, organism="hsa"): url = f"https://rest.kegg.jp/link/{organism}/{pathway_id}" response = requests.get(url) if response.status_code == 200: genes = [] for line in response.text.strip().split("\\n"): parts = line.split("\\t") if len(parts) == 2: genes.append(parts[1].replace(f"{organism}:", "")) return genes return [] # Example genes = get_pathway_genes("hsa04110") print(f"Genes: {len(genes)}") print(genes[:10]) """) print("\n=== 3. Map Genes to Pathways ===") print(""" def map_gene_to_pathways(gene_id, organism="hsa"): url = f"https://rest.kegg.jp/link/pathway/{organism}:{gene_id}" response = requests.get(url) if response.status_code == 200: pathways = [] for line in response.text.strip().split("\\n"): parts = line.split("\\t") if len(parts) == 2: pathways.append(parts[1]) return pathways return [] # Example pathways = map_gene_to_pathways("7157") # TP53 print(pathways) """) print("\n=== 4. Simple Enrichment ===") print(""" import requests from scipy.stats import hypergeom import numpy as np def kegg_enrichment(gene_list, background, organism="hsa"): \"\"\"Simple KEGG enrichment.\"\"\" # Get all pathways url = f"https://rest.kegg.jp/list/pathway/{organism}" response = requests.get(url) pathways = {} for line in response.text.strip().split("\\n"): parts = line.split("\\t") if len(parts) == 2: pathways[parts[0]] = parts[1] # For each pathway results = [] N = len(background) n = len(gene_list) for pathway_id, name in pathways.items(): # Get genes in pathway genes = get_pathway_genes(pathway_id, organism) K = len(genes) k = len(set(gene_list) & set(genes)) if k > 0: p = hypergeom.sf(k-1, N, K, n) results.append({ "pathway": pathway_id, "name": name, "count": k, "pvalue": p }) # Sort results.sort(key=lambda x: x["pvalue"]) return results # Example # results = kegg_enrichment(gene_list, background) """) print("\n=== 5. GSEA with GSEApy ===") print(""" import gseapy # GSEA gsea = gseapy.gsea( data="expression_data.txt", gene_sets="KEGG_2021_Human", cls="phenotype.cls", permutation_num=1000, outdir="gsea_output" ) # View print(gsea.res2d) """) print("\n=== 6. Enrichr with GSEApy ===") print(""" import gseapy # Enrichr enr = gseapy.enrichr( gene_list=gene_list, gene_sets="KEGG_2021_Human", organism="human", outdir="enrichr_output" ) # View print(enr.results) """) print("\n=== 7. Visualization ===") print(""" import matplotlib.pyplot as plt import numpy as np def plot_enrichment(results, output="enrichment.png"): \"\"\"Plot enrichment results.\"\"\" results = results[:20] names = [r["name"] for r in results] pvals = [-np.log10(r["pvalue"]) for r in results] plt.figure(figsize=(10, 8)) plt.barh(names, pvals, color='purple') plt.xlabel("-log10(p-value)") plt.title("KEGG Pathway Enrichment") plt.tight_layout() plt.savefig(output) plt.show() # Use # plot_enrichment(results) """) print("\n=== 8. Complete Pipeline ===") print(""" import requests from scipy.stats import hypergeom import pandas as pd def kegg_pipeline(gene_list, background, organism="hsa"): \"\"\"Complete KEGG enrichment pipeline.\"\"\" # Get all pathways url = f"https://rest.kegg.jp/list/pathway/{organism}" response = requests.get(url) pathways = {} for line in response.text.strip().split("\\n"): parts = line.split("\\t") if len(parts) == 2: pathways[parts[0]] = parts[1] # Enrichment results = [] N = len(background) n = len(gene_list) for pathway_id, name in pathways.items(): genes = get_pathway_genes(pathway_id, organism) K = len(genes) k = len(set(gene_list) & set(genes)) if k > 0: p = hypergeom.sf(k-1, N, K, n) results.append({ "pathway": pathway_id, "name": name, "count": k, "pvalue": p }) # Sort and FDR results.sort(key=lambda x: x["pvalue"]) df = pd.DataFrame(results) df["fdr"] = df["pvalue"] * len(df) / (df.index + 1) df = df[df["fdr"] < 0.05] return df # Use # df = kegg_pipeline(gene_list, background) # print(df.head()) """)

⚖️ Database Comparison

🎯 Which Pathway Database?

KEGG is not the only pathway database. Reactome, WikiPathways, and BioCyc offer different strengths.

Comparison

KEGG: Comprehensive. Reactome: Curated. WikiPathways: Community. BioCyc: Many organisms. PANTHER: Classification. Pathway Commons: Integration.

⚖️ KEGG⚖️ Reactome⚖️ WikiPathways⚖️ BioCyc
Database Coverage Curation Species Best For
KEGGComprehensiveManual8,000+General, metabolic
ReactomeDetailedExpert~20Human, signaling
WikiPathwaysCommunityCommunity~30Emerging topics
BioCycManyCurated~20,000Microbes, metabolic
PANTHERClassificationCurated~100Protein families
Pathway CommonsIntegratedIntegrated~20Integration
MetaCycMetabolicCurated~3,000Metabolism
Ingenuity (IPA)CommercialCuratedHuman, mouse, ratCommercial
# Decision Guide print("=== Choose Database by Need ===") print(""" General pathway analysis: → KEGG Human signaling: → Reactome Emerging topics: → WikiPathways Microbes: → BioCyc Metabolism: → MetaCyc Integration: → Pathway Commons Commercial: → IPA """) print("\n=== Speed Comparison ===") print(f"{'Database':<20} {'Pathways':<15} {'Species':<15} {'API':<10}") print("-" * 60) print(f"{'KEGG':<20} {'~500':<15} {'8,000+':<15} {'Yes':<10}") print(f"{'Reactome':<20} {'~2,500':<15} {'~20':<15} {'Yes':<10}") print(f"{'WikiPathways':<20} {'~3,000':<15} {'~30':<15} {'Yes':<10}") print(f"{'BioCyc':<20} {'~3,000':<15} {'~20,000':<15} {'Yes':<10}") print("\n=== When to Use Which ===") print(""" USE KEGG: ✓ General ✓ Metabolic ✓ Cross-species ✓ Enrichment USE Reactome: ✓ Human ✓ Signaling ✓ Detailed ✓ Curated USE WikiPathways: ✓ Community ✓ Emerging ✓ Open USE BioCyc: ✓ Microbes ✓ Metabolic ✓ Many organisms USE PANTHER: ✓ Classification ✓ Protein families """)

💼 Use Cases

🎯 Where KEGG Is Used

KEGG pathway analysis is used across biology: transcriptomics, proteomics, metabolomics, disease research, and drug discovery.

Use Cases

Transcriptomics: DE genes. Proteomics: Proteins. Metabolomics: Metabolites. Disease: Mechanisms. Drug: Targets. Microbiome: Functions.

💼 Transcriptomics💼 Proteomics💼 Metabolomics💼 Disease
# Use Cases print("=== 1. Transcriptomics ===") print(""" Use case: - RNA-seq - DE genes - Pathway enrichment Workflow: 1. RNA-seq 2. DE analysis 3. KEGG enrichment 4. Interpretation Example: Cancer vs normal → DE genes → Cell cycle enriched → Proliferation """) print("\n=== 2. Proteomics ===") print(""" Use case: - Mass spectrometry - Protein list - Pathway enrichment Workflow: 1. Proteomics 2. Protein list 3. KEGG enrichment 4. Interpretation Example: Drug treatment → Proteins → Apoptosis enriched → Mechanism """) print("\n=== 3. Metabolomics ===") print(""" Use case: - Metabolites - Metabolic pathways - KEGG enrichment Workflow: 1. Metabolomics 2. Metabolite list 3. KEGG enrichment 4. Interpretation Example: Disease → Metabolites → TCA cycle enriched → Mechanism """) print("\n=== 4. Multi-Omics ===") print(""" Use case: - RNA + protein + metabolite - Integrated analysis - Pathways Workflow: 1. Multi-omics 2. Integration 3. KEGG enrichment 4. Systems view Example: Cancer → Multi-omics → Pathways → Mechanism """) print("\n=== 5. Disease Analysis ===") print(""" Use case: - Disease genes - Pathways - Mechanisms Workflow: 1. Disease genes 2. KEGG enrichment 3. Pathways 4. Mechanism Example: Alzheimer's → Disease genes → Pathways → Amyloid """) print("\n=== 6. Drug Discovery ===") print(""" Use case: - Drug targets - Pathways - Repurposing Workflow: 1. Drug targets 2. KEGG enrichment 3. Pathways 4. Repurposing Example: Drug → Targets → Pathways → New uses """) print("\n=== 7. Microbiome ===") print(""" Use case: - Metagenomics - KEGG modules - Functions Workflow: 1. Metagenomics 2. KO annotation 3. KEGG modules 4. Functions Example: Gut microbiome → KO → Pathways → Functions """) print("\n=== 8. Single-Cell ===") print(""" Use case: - scRNA-seq - Cell types - Pathways Workflow: 1. scRNA-seq 2. Clustering 3. KEGG enrichment 4. Cell-type specific Example: Tumor → Cell types → Pathways → Heterogeneity """) print("\n=== 9. Plant Biology ===") print(""" Use case: - Plant genomics - Pathways - Traits Workflow: 1. Plant genomics 2. DE genes 3. KEGG enrichment 4. Traits Example: Drought → DE genes → Pathways → Tolerance """) print("\n=== 10. Microbiology ===") print(""" Use case: - Bacterial genomes - Metabolic pathways - Functions Workflow: 1. Bacterial genome 2. KO annotation 3. KEGG pathways 4. Functions Example: Pathogen → Genome → Pathways → Targets """)

❓ FAQ

🎯 Frequently Asked Questions

Answers to common questions about KEGG and pathway analysis.

What is KEGG?

KEGG (Kyoto Encyclopedia of Genes and Genomes) is a comprehensive database resource for understanding high-level functions and utilities of biological systems from molecular-level information. It includes pathway maps, gene catalogs, chemical compounds, and disease information.

What is a KEGG pathway?

A KEGG pathway is a diagrammatic representation of molecular interactions, reactions, and networks in a biological process. Examples include metabolic pathways (e.g., glycolysis, TCA cycle), signaling pathways (e.g., MAPK, p53), and disease pathways. Each pathway has a unique ID (e.g., hsa00010 for glycolysis).

What is a KO number?

A KO (KEGG Orthology) number is a unique identifier (e.g., K00001) assigned to a group of orthologous genes that share the same function. KO numbers link genes across species to KEGG pathways, enabling cross-species pathway analysis.

What is pathway enrichment analysis?

Pathway enrichment analysis identifies KEGG pathways that are over-represented (enriched) in a set of genes (e.g., differentially expressed genes) compared to a background. It uses statistical tests (hypergeometric, Fisher's exact) and corrects for multiple testing (FDR). Tools include clusterProfiler, KEGGREST, and DAVID.

What is GSEA?

GSEA (Gene Set Enrichment Analysis) is a method that determines whether a defined set of genes (e.g., a KEGG pathway) shows statistically significant, concordant differences between two biological states. Unlike enrichment, GSEA uses all genes ranked by expression, not just a cutoff list.

What is the difference between KEGG and Reactome?

KEGG is a comprehensive database covering metabolism, signaling, and disease across thousands of species. Reactome is a curated, expert-written database of human and model organism pathways, with detailed molecular reactions. KEGG is broader; Reactome is deeper for human.

How do I map my genes to KEGG pathways?

Use KEGG Mapper (web), clusterProfiler (R), KEGGREST (R), or the KEGG API (Python). You can map gene IDs to KO numbers, then to pathways. For cross-species, use KO numbers.

What is the best background for KEGG enrichment?

The best background is the set of all genes that could have been detected in your experiment. For RNA-seq, use all expressed genes. For microarrays, use all probes. For genome-wide, use all genes. The choice of background significantly affects results.

What is a pathway map?

A pathway map is a diagram showing the relationships between genes, proteins, and metabolites in a biological process. KEGG pathway maps are manually drawn and use standard symbols (e.g., rectangles for genes, circles for metabolites).

Can I use KEGG for metagenomics?

Yes. KEGG is widely used in metagenomics. You can annotate genes with KO numbers using tools like GhostKOALA or KofamKOALA, then map to KEGG modules and pathways to infer community functions.

# Quick Reference print("=== Quick Answers ===") print(""" Q: What is KEGG? A: Pathway database. Q: KO number? A: Ortholog group ID (K00001). Q: Pathway ID? A: Organism + number (hsa00010). Q: Enrichment? A: Over-represented pathways. Q: GSEA? A: Ranked gene list analysis. Q: Tools? A: clusterProfiler, pathview, KEGGREST, GSEApy. Q: R? A: clusterProfiler, pathview, KEGGREST. Q: Python? A: GSEApy, requests. Q: Web? A: DAVID, KEGG Mapper, g:Profiler. Q: Cross-species? A: Use KO numbers. Q: Free? A: Yes — all major tools are free. """)

🚀 Future

🎯 Where Pathway Analysis Is Going

The future is AI-integrated, single-cell, spatial, multi-omics, and knowledge graphs.

Future Directions

AI: Integration. Single-cell: Resolution. Spatial: Location. Multi-omics: Integrated. Graphs: Knowledge. Real-time: Updates.

🚀 AI🚀 Single-cell🚀 Spatial🚀 Multi-omics
# Future of Pathway Analysis print("=== 1. AI Integration ===") print(""" Trend: AI + pathways Ideas: - AI-powered annotation - Predictive pathways - NLP for literature - Knowledge graphs Examples: - AlphaFold - GPT for biology - AI curation Future: - Automated - Predictive - Integrated """) print("\n=== 2. Single-Cell ===") print(""" Trend: Single-cell pathways Current: - Bulk pathways - Average Future: - Single-cell pathways - Cell-type specific - Trajectory Examples: - scRNA-seq + KEGG - Cell atlas - Developmental Future: - Cell-type resolution - Trajectories - Dynamics """) print("\n=== 3. Spatial ===") print(""" Trend: Spatial pathways Current: - Tissue-level Future: - Spatial pathways - Location-specific - Microenvironment Examples: - Spatial transcriptomics - Imaging - Tumor Future: - Spatial resolution - Microenvironment - Context """) print("\n=== 4. Multi-Omics ===") print(""" Trend: Multi-omics pathways Current: - Single omics Future: - Multi-omics pathways - Integration - Systems Examples: - RNA + protein + metabolite - Pathway analysis - Integrated view Future: - Unified - Integrated - Systems-level """) print("\n=== 5. Knowledge Graphs ===") print(""" Trend: Knowledge graphs Current: - Separate databases - Flat files Future: - Knowledge graphs - Linked data - Reasoning Examples: - KEGG + Reactome - Wikidata - SPOKE Future: - Unified - Reasoning - Discovery """) print("\n=== 6. Real-Time ===") print(""" Trend: Real-time pathways Current: - Periodic updates - Static Future: - Real-time - Streaming - Live Examples: - COVID-19 - Outbreak - Surveillance Future: - Real-time - Continuous - Dynamic """) # TIMELINE print("\n=== Timeline ===") print(""" Past: 1995: KEGG founded 2000: KEGG PATHWAY 2005: KO, BRITE 2010: Enrichment tools 2015: GSEA, clusterProfiler 2020: Multi-omics Present: 2025: AI integration 2026: Single-cell Future: 2030: Spatial 2035: Knowledge graphs 2040: Real-time? """) print("\n=== Final Thoughts ===") print(""" KEGG & pathway analysis is: - Essential for omics - Comprehensive - Cross-species - Evolving The future is: - AI-integrated - Single-cell - Spatial - Multi-omics - Knowledge graphs - Real-time "Pathways are the language of biology." Keep exploring. Keep building. """)