What is KEGG?

A short guide to KEGG pathways, enrichment analysis, and responsible interpretation of methylation-associated gene lists

A map, not a measurement

The Kyoto Encyclopedia of Genes and Genomes (KEGG) organizes genes, proteins, metabolites, diseases, and drugs into manually drawn pathway maps. It helps researchers ask whether a result list contains more members of a known biological system than expected by chance. KEGG does not measure pathway activity and does not prove that a pathway causes disease.

A pathway connects receptors, signaling proteins, gene regulation, immune responses, and epithelial outcomes. This original illustration is conceptual rather than a KEGG map.

What a KEGG pathway contains

A KEGG PATHWAY map is a network of molecular interactions, reactions, and processes. Broad collections include metabolism, genetic and environmental information processing, cellular processes, organismal systems, human diseases, and drug development. Organism-specific identifiers connect human genes to these maps.

flowchart LR
  D["FDR-controlled<br/>DMRs"] --> A["Map regions to genes<br/>with explicit rules"]
  A --> L["Candidate gene list"]
  U["Genes linked to all<br/>regions actually tested"] --> E["KEGG enrichment test"]
  L --> E
  E --> R["Pathway overlap<br/>effect + adjusted p-value"]
  R --> I["Biological hypothesis<br/>for validation"]

How enrichment works

Over-representation analysis compares two proportions: the fraction of candidate genes belonging to a pathway and the fraction in a relevant background universe. A hypergeometric or Fisher exact test is common, followed by false-discovery-rate correction across pathways. Ranked-list methods such as gene-set enrichment analysis can instead test whether a pathway’s genes accumulate toward one end of an ordered statistic.

The background matters enormously. For this study it should represent genes that could have been discovered after coverage, filtering, and region-to-gene mapping—not every human gene. Using the whole genome can make enrichment appear stronger simply because many genes were never testable.

How KEGG helps this study

After coverage-aware DML/DMR analysis, KEGG can organize candidate genes into interpretable systems potentially related to:

• inflammatory and immune signaling;
• epithelial integrity and cell adhesion;
• apoptosis, cell cycle, and DNA repair;
• host–microbe interactions; and
• colorectal tumor biology.

These labels are prompts for biological follow-up. A methylated promoter near a pathway gene does not automatically mean the gene is less expressed; the effect depends on genomic context, cell mixture, direction, and functional evidence.

What should be reported

Item	Why readers need it
KEGG release or access date	Pathway membership changes over time
Genome build and identifier conversion	Prevents silent coordinate/gene mismatches
CpG/DMR-to-gene rule	One region may map to several genes—or none
Tested-feature background	Defines the actual null comparison
Overlap genes and counts	Shows what produced the pathway result
Effect measure and adjusted p-value	Separates magnitude from multiplicity
Direction of methylation change	Avoids mixing hyper- and hypomethylated biology

In R, clusterProfiler provides transparent KEGG enrichment workflows. Reproducibility still depends on saving identifiers, inputs, parameters, software version, and database date. Users should also check KEGG’s database and licensing information for the intended access and reuse.

ImportantBottom line

KEGG converts a statistically supported gene list into a structured biological question. It is strongest when the gene-mapping rule, tested background, overlap, and uncertainty are visible—and weakest when a colorful pathway name is treated as proof of mechanism.

Compare Reactome → · See the statistical workflow → · Read the software guide →