KEGGgraph

KEGGgraph is an R/Bioconductor package that provides an interface between the KEGG PATHWAY database and R's graph object model, together with a collection of functions to parse, dissect, analyze, and visualize pathway networks. It was written by Jitao David Zhang (with contributions from Paul Shannon and Herve Pages), is distributed under GPL (>= 2), and has been part of Bioconductor since release 2.4, making it one of the longer-lived pathway-analysis packages in the ecosystem (version 1.72.0 in Bioconductor 3.23).

The package's core job is parsing KGML, the KEGG Markup Language XML export used to describe KEGG maps (KGML, produced by KEGG). The function parseKGML reads a KGML file into a KEGGPathway object that retains pathway metadata such as title, organism, and link; parseKGML2Graph converts that into a graphNEL object with directed edges, the native graph class of the Bioconductor graph package. A parseKGML2DataFrame helper emits a simple edge-list-style data frame for users who want the relations without the graph machinery. Node identity is controlled by the expandGenes flag: with expandGenes=TRUE nodes carry unique KEGG identifiers of the form org:xxxx (for example hsa:1432), whereas with expandGenes=FALSE KEGG's internal auto-increment entry indices are used, which produces a simplified graph but yields duplicated node names that prevent multiple pathway graphs from being merged cleanly.

In the graph-data and format-conversion landscape, KEGGgraph acts as a bridge from KEGG's domain-specific XML into general-purpose graph tooling. Because its output is a standard graphNEL, results plug directly into the surrounding Bioconductor graph stack — RBGL for graph algorithms and Rgraphviz (Graphviz bindings) for layout and rendering — and can be exported for use in Cytoscape, igraph, NetworkX, or graph-tool. It is a dependency of downstream packages such as pathview, EnrichmentBrowser, and DEGraph. It overlaps in scope with other KGML converters like KEGGconverter, KEGGtranslator, and kgml2sif, but is distinguished by its tight, idiomatic integration with R rather than by producing standalone interchange files.

Its main limitations follow from being KGML-specific and R-centric: it only handles KEGG's own XML, the expandGenes trade-off forces a choice between merge-safe uniqueness and pathway simplicity, and querying live pathways still relies on KEGG's network services and licensing terms. For R users doing pathway topology or enrichment analysis, however, it remains a compact and dependable entry point.