The KEGG database contains three main components for genome/metagenome annotation:

• the collection of internally annotated gene catalogs for sequenced genomes, both for cellular organisms (called KEGG organisms) and viruses, and additional publication-based protein sequences in the GENES database,
• the knowledge base of high-level functions represented as molecular interaction, reaction and relation networks in the PATHWAY, BRITE and MODULE databases, and
• the knowledge base of molecular-level functions associated with ortholog groups in the KO database, where most KO entries are defined in a context-dependent manner as nodes of the KEGG molecular networks.

In general, KO entries (identified by K numbers) also represent sequence similarity groups. Thus, the sequence similarity search of a query genome against KEGG GENES is a search for most appropriate K numbers. The assigned set of K numbers can then be used to reconstruct KEGG pathway maps, BRITE hierarchies and KEGG modules, enabling interpretation of high-level functions.

Automatic KO assignment can be done by the following web servers

• BlastKOALA
• GhostKOALA

or by an interactive tool of

• Assign KO

and the result of KO assignment is linked to KEGG Mapper Reconstruct tool.

All genome pairs in KEGG are subject to SSDB (Sequence Similarity DataBase) computation using the SSEARCH program for both amino acid sequences (protein coding genes) and nucleotide sequences (RNA genes). For each gene, an organism-based list of similarity neighbors is generated and displayed in a tabular form, called GFIT table. It shows the best hit sequence in each matching organism, as can be viewed from GFIT button in the GENES entry page. The collection of GFIT tables is the basis for both manual and automatic annotation of the GENES database.

The KEGG Annotation Guide is a collection of HTML tables, called BRITE tables, showing summary views of the current annotation of the GENES database, such as how K numbers are defined and assigned for distinguishing related genes and for comparing different subunit structures.

Comparing subunit structures or gene sets

• 01610 Ribosomal proteins
• 01611 RNA polymerases
• 01632 DNA polymerases
• 01610 Aminoacyl-tRNA synthetases

Distinguishing related genes

• 01601 Enzymes of 2-oxocarboxylic acid metabolism
• 01602 Dioxygenases

Another set of BRITE tables contains signature KOs and/or signature modules, which can be used to infer phenotypic features of organisms.

Metabolic capacity

• 01620 Photosynthetic and chemosynthetic capacities

Pathogenicity and drug resistance

• 01600 Antimicrobial resistance: KEGG signatures