Paper 9: Enrichment Map - bcb420-2025/Keren_Zhang GitHub Wiki

Table of Contents Pathway enrichment analysis and visualization of omics data using g:Profiler, GSEA, Cytoscape and EnrichmentMap Key Concepts Pathway Enrichment Analysis Major Tools Used Protocol Steps Stage 1: Define Gene List Stage 2: Pathway Enrichment Analysis Option A: g:Profiler Option B: GSEA Stage 3: Visualization with EnrichmentMap Advantages Limitations Example Results Implementation Notes References

DOI: 10.1038/s41596-018-0103-9

This protocol provides a step-by-step guide for pathway enrichment analysis of omics data, covering three major stages: 1. Definition of a gene list from omics data 2. Pathway enrichment analysis using statistical methods 3. Visualization and interpretation of results

• Identifies biological pathways enriched in a gene list more than expected by chance
• Helps interpret gene lists from RNA-seq, genome-sequencing, and other omics experiments
• Projects disease data onto known mechanisms to increase statistical and interpretative power

• g:Profiler: For analysis of gene lists using Fisher's exact test
• GSEA (Gene Set Enrichment Analysis): For analysis of ranked gene lists
• Cytoscape with EnrichmentMap: For visualization of pathway enrichment results

• Input can be:

  ** Unranked gene list (e.g., mutated genes) → use g:Profiler  
  ** Ranked gene list (e.g., differentially expressed genes) → use GSEA  

• Example data provided:

  ** Cancer driver genes (for g:Profiler)  
  ** Ovarian cancer subtypes (for GSEA)  

• Uses Fisher's exact test for gene lists
• Key parameters:

  ** Ordered query option for ranked lists  
  ** Size filters (recommended: 5-350 genes per pathway)  
  ** Multiple testing correction (FDR Q < 0.05)  

• Analyzes whole-genome ranked lists without thresholding
• Key parameters:

  ** Number of permutations (default: 1,000)  
  ** Gene set size limits (recommended: max 200 genes)  
  ** Phenotype labels (pos/neg for up/down-regulated)  

• Creates network of enriched pathways where:

  ** Nodes = pathways  
  ** Edges = shared genes between pathways  

• Features:

  ** Color by enrichment score (red/blue for up/down)  
  ** Cluster similar pathways into biological themes  
  ** Heatmaps show gene expression patterns  

• Improves statistical power by aggregating gene-level signals
• Results are easier to interpret than raw gene lists
• Facilitates integration of diverse omics data types
• Uses freely available, frequently updated software

• Less effective for pathways regulated by few genes
• Pathway boundaries can be arbitrary across databases
• Biased toward well-annotated pathways
• May miss "dark matter" genes without pathway annotations

• g:Profiler analysis of cancer driver genes identified pathways like:

  ** "Positive regulation of Ras protein signal transduction"  
  ** "Regulation of interferon-gamma-mediated signaling"  

• GSEA analysis of ovarian cancer showed:

  ** Mesenchymal subtype enriched for cell cycle pathways  
  ** Immunoreactive subtype enriched for immune response pathways  

• Complete protocol takes ~4.5 hours
• Designed for biologists with no bioinformatics training
• Includes troubleshooting guide for common issues
• Provides supplementary protocols for advanced analyses

• g:Profiler: Reimand et al. (2016) Nucleic Acids Res
• GSEA: Subramanian et al. (2005) PNAS
• EnrichmentMap: Merico et al. (2010) PLoS ONE