Functional annotation at the metagenomic scale
eggNOG-mapper annotates large sequence sets — metagenomic protein catalogs, whole proteomes, CDS or prokaryotic contigs — adjusting the taxonomic context individually for every query, using eggNOG 7 orthologous groups and their gene trees.
What's new in v3
eggNOG-mapper v3 rebuilds the annotation engine around eggNOG 7: a larger curated foundation (12,535 reference proteomes, 59.3M proteins, 3.18M orthologous groups), per-query taxonomic scoping, and a cascade transfer that keeps only the closest curated evidence.
eggNOG v7 database
A larger, curated orthology foundation: 12,535 reference proteomes — 10,756 Bacteria, 457 Archaea and 1,322 Eukaryota — with 59.3M proteins in 3.18M orthologous groups.
Curated-only donors
Functional terms are transferred only from manually reviewed sources — no electronically inferred annotations are propagated.
Updated functional sources
Terms are drawn from current releases of Swiss-Prot, KEGG, Gene Ontology, COG, PFAM and EC.
Per-seed taxonomic ceiling
The taxonomic scope is adjusted for every query, so mixed metagenomic communities annotate correctly without one global clade.
Cascade annotation engine
A lazy closest-cascade walks outward tier by tier and keeps the nearest evolutionary evidence available.
How to cite
If you use results from this service, please cite both the tool and the database.
Cantalapiedra CP, Hernández-Plaza A, Letunic I, Bork P, Huerta-Cepas J. eggNOG-mapper v2. Mol Biol Evol. 2021;38(12):5825–5829. doi:10.1093/molbev/msab293
Hernández-Plaza A, Deng D, Robledo-Yagüe A, et al. eggNOG v7. Nucleic Acids Res. 2026;54(D1):D402–D408. doi:10.1093/nar/gkaf1249
Functional annotation was
performed with eggNOG-mapper v3.0.0 (Cantalapiedra et al., 2021) using orthology assignments from the
eggNOG 7 database (Hernández-Plaza et al., 2026). Homology searches used DIAMOND in sensitive mode
(e-value ≤ 0.001), transferring GO, KEGG, COG and PFAM terms from orthologs within the automatically
adjusted taxonomic scope (--tax_scope auto).
Open source under AGPL-3.0. Report issues and ask questions on GitHub.