Skip to content

The KEGG pathways database

KEGG is the pathway database every PaintOmics job runs against. Its maps are drawn by hand and published together with the coordinates of everything on them, which is what makes a KEGG pathway paintable: PaintOmics puts your values into the boxes KEGG already drew.

The application describes it in Step 2, in the Multiple databases used card that appears when a job runs more than one database, as a database resource for understanding high-level functions and utilities of the biological system, such as the cell, the organism and the ecosystem, from molecular-level information — see kegg.jp/kegg.

KEGG is not optional

In Step 1 the KEGG checkbox is ticked, greyed out and tagged required, and the server adds KEGG to the job's database list whatever the form posted. Every job therefore has KEGG pathways, even one run for the sake of Reactome or MapMan. The other three databases are offered only when this server has installed them for the organism you chose.

What PaintOmics installs from KEGG

Installing an organism pulls two kinds of data.

Per organism:

  • the list of that organism's pathways, and one KGML file per pathway — the machine-readable form of the map, which carries the position, width and height of every gene and compound box;
  • the gene identifier tables the identifier translation works through;
  • a pathway-to-pathway network, drawn in Step 3.

Shared by every organism on the server:

  • the reference pathway images (map00230.png and so on) and their thumbnails;
  • the pathway classification, KEGG BRITE br08901;
  • the compound list with every synonym KEGG records, the pathway-to-compound table, and a ChEBI-to-KEGG compound mapping.

The shared download is about 1.4 GB and each species adds roughly 200–400 MB. See Installing PaintOmics.

One further piece of KEGG is bundled with the application rather than downloaded: the chemical classification of compounds, KEGG BRITE br08001, which is what the metabolite class activity analysis groups by.

Why KEGG pathways can be painted

KEGG publishes both the picture and the geometry. The picture is the reference map — one image per pathway, the same for every organism — and the geometry comes from your organism's own KGML. When you paint mmu00230, PaintOmics serves the image map00230.png and reads the box coordinates out of the mouse KGML, so the diagram is KEGG's own drawing with your data laid over it, one box per matched feature and one coloured cell per condition.

This is the same mechanism Reactome uses, from a downloaded PNG. MapMan works the same way from its own diagrams. OmniPath has no diagram at all and opens as an interaction network instead.

The per-database match counts in Step 2

Step 2 counts, for each omic, how many of your features carry an identifier each database is keyed on. Reading across a row is how you find out why one database matched more of your data than another.

Identifiers

KEGG pathways name their genes with the organism's own KEGG gene identifiers, and PaintOmics translates whatever you uploaded into the identifier space that organism's data is keyed on. That target is per organism, not universal. The shipped configuration names 23 organisms and splits them in two.

  • Ten resolve through NCBI Gene IDs: human, mouse, rat, cow, chimpanzee, green anole, chicken, Xenopus, pig and dog.
  • The other thirteen resolve through KEGG's own gene ids: zebrafish, fly, worm, budding yeast, S. pombe, P. falciparum, D. discoideum, Bifidobacterium animalis subsp. lactis BB-12, tomato, potato, Arabidopsis, and rice under both of the KEGG codes it has. Those two codes spell a rice gene differently: KEGG names osa genes by their NCBI Gene ID, and dosa genes by their RAP-DB locus (Os01g0147900). They are separate installations, so pick the one whose identifiers your file uses.

An organism with no entry in the shipped configuration falls back to KEGG gene ids and KEGG gene symbols.

Metabolites are matched against KEGG compound names, case-insensitively. Bare KEGG compound ids (C00002) work because the installer files each id as a name of itself, and ChEBI ids are accepted with or without the chebi: prefix. A name that matches several KEGG compounds raises the disambiguation cards in Step 2.

Supported identifiers has the per-organism detail.

Classification

Each KEGG pathway carries two classification levels taken from br08901: a top level — Metabolism, Genetic Information Processing, Environmental Information Processing, Cellular Processes, Organismal Systems, Human Diseases, Drug Development — and a second level beneath it. Step 3 draws that as the Category Distribution pie and the Filter by category tree; only classes that have pathways for your organism appear. See Pathway classification.

Two analyses that are KEGG-only

Two things in Step 3 come from KEGG data and are unaffected by which other databases the job runs:

How KEGG compares with the others

KEGG Reactome MapMan OmniPath
Offered for every installed organism 15 organisms in the shipped configuration 4 plant organisms human, mouse, rat
Can be switched off no yes yes yes
Pathway opens as a painted KEGG map a painted Reactome diagram a painted MapMan diagram an interactive interaction network
Genes keyed on the organism's KEGG gene ids Reactome gene ids MapMan gene ids UniProt accessions
Matches metabolites yes yes, through ChEBI yes, through MapMan bins no
Classification KEGG BRITE br08901, two levels Reactome's event hierarchy, two levels MapMan's own, two levels assigned by PaintOmics, resource second

Which of the four you can actually tick is a fact about the server you are using, not a fixed list. The Step 1 checkboxes are the authority: a database this deployment has not installed for your organism is disabled and labelled not installed.

Note

KEGG data is dated per organism. The administration panel shows the KEGG information date and the mapping date for every installed organism, and how many pathways your organism has is reported per job in the Step 3 summary band — not by any number quoted in this documentation.