Workflow of evaluation of the potential functionomes. Detailed workflow of the three annotation servers, KAAS, MEGAN4, and MG-RAST using query sequences after gene finding process of sequenced data; KAAS and MEGAN4 use BLASTP and BLASTX for amino acid and nucleotide query sequences respectively and the MG-RAST uses only BLASTX. All use different databases, i.e. KEGG GENES for KAAS, NCBI-NR for MEGAN4, and M5nr  for MG-RAST (M5nr includes the SEED as a subset.), and different default threshold values for the BLAST hits. Each server converts the hit entries to the corresponding orthology IDs for functional annotation and pathway/module/subsystem mapping. Red colored texts of KAAS indicate its improvements in the current study (see Assignment of the query sequences to KO identifiers in the Methods section).
Takami et al. BMC Genomics 2012 13:699 doi:10.1186/1471-2164-13-699