eggNOG-mapper v3

eggNOG-mapper is a tool for fast functional annotation of novel sequences (proteins, CDS, genomes, metagenomes) using precomputed orthology data from the eggNOG database. Functional terms — GO, KEGG orthologs/pathways/modules, EC numbers, PFAM domains, CAZy families, BiGG reactions and more — are transferred from fine-grained orthologs, which is more precise than best-hit/BLAST transfer because it avoids annotation from close paralogs.

v3 targets the eggNOG 7 database (~59 million proteins across the tree of life). It is not compatible with eggNOG 5 data. See USAGE.md for the complete manual (basic → advanced → HPC → containers).

Three ways to run

	Best for	Install
① Free web service — https://eggnog-mapper.cgmlab.org	small/moderate jobs, zero setup	none
② Apptainer/Singularity image	reproducible runs, HPC clusters	one .sif (all tools bundled)
③ pip / from source	pipelines, development	Python package + databases

Prebuilt images and databases: https://data.cgmlab.org/eggnog-mapper/

What's new in v3

• eggNOG 7 database with integer-encoded orthology and phylogeny-aware speciation events (~59M proteins). eggNOG 5 is no longer supported.
• Curated-only functional donors — only manually curated terms are used as annotation donors, stopping the propagation of automated misannotations, while achieving better coverage than v2.
• Per-seed taxonomic ceiling (--tax_scope auto, default) replaces the old predefined scope lists; each seed narrows to its most informative level. Fixed clades (Metazoa, 33208, …) are still accepted.
• Cascade annotation engine with a lazy closest-cascade on by default and seed sorting + de-duplication — large speedups on redundant proteome/UniProt scale inputs, byte-identical output.
• Compact 22-column output with a positional annotation_confidence field (per-source confidence, documented in the header). See USAGE › Output files.
• --selftest — one command downloads a tiny reference set and verifies the install reproduces the expected annotations.
• Self-contained Apptainer image bundling DIAMOND, MMseqs2, HMMER and Prodigal (apptainer/build.sh).
• Optional MMseqs2 backend, gzip/bzip2 input autodetection, Cython-accelerated inner loops, and --resume for interrupted runs.

Install

pip install eggnog-mapper
# or from source:
git clone https://github.com/eggnogdb/eggnog-mapper.git
cd eggnog-mapper && pip install .

Requires Python ≥ 3.9 and the search tools on PATH:

tool	role	in v3
DIAMOND	seed search	default backend
MMseqs2	seed search	optional backend (-m mmseqs)
Prodigal	gene prediction	genome/metagenome input
HMMER	domain realignment	only for --pfam_realign

Install e.g. conda install -c bioconda diamond mmseqs2 prodigal hmmer. All of these are bundled in the Apptainer image — nothing to install there.

Databases

python download_eggnog_data.py -y --data_dir /path/to/data

Data is versioned by MAJOR.MINOR (emapper-3.0/) and pinned to the eggNOG DB version: a minor bump (3.0 → 3.1) means the DB changed and must be re-downloaded; patch releases reuse the same data. Core download ≈ 45 GB (annotation DB + DIAMOND DB + taxonomy + prebuilt caches); add -M for the optional MMseqs2 DB, -P for Pfam. Details: USAGE › Databases.

Quick start

# Proteins (default input type)
emapper.py -i proteins.fa -o my_run --data_dir /path/to/data --cpu 8

# CDS / genome / metagenome
emapper.py -i cds.fna       --itype CDS        -o my_run --data_dir DATA --cpu 8
emapper.py -i genome.fna    --itype genome     -o my_run --data_dir DATA --cpu 8
emapper.py -i contigs.fna   --itype metagenome -o my_run --data_dir DATA --cpu 8

# Verify your installation end-to-end
emapper.py --selftest

Output is written as my_run.emapper.annotations (22-column TSV) plus optional .xlsx / .orthologs / .gff. Full options, output-column reference, advanced usage and HPC performance tuning are in USAGE.md.

Documentation

USAGE.md — the complete manual: basic usage, databases, output format, Apptainer, advanced options, and maximum-performance tips for HPC clusters.

Citation

If you use eggNOG-mapper, please cite:

[1] eggNOG-mapper v2: functional annotation, orthology assignments, and domain
    prediction at the metagenomic scale. Carlos P. Cantalapiedra,
    Ana Hernandez-Plaza, Ivica Letunic, Peer Bork, Jaime Huerta-Cepas. 2021.
    Molecular Biology and Evolution, msab293, https://doi.org/10.1093/molbev/msab293

[2] eggNOG v7: phylogeny-based orthology predictions and functional annotations.
    Ana Hernández-Plaza, Ziqi Deng, Fabian Robledo-Yagüe, Damian Szklarczyk,
    Christian von Mering, Peer Bork, Jaime Huerta-Cepas. Nucleic Acids Research,
    Volume 54, Issue D1, 6 January 2026, Pages D402-D408.
    https://doi.org/10.1093/nar/gkaf1249

Please also cite the search tool used (the end-of-run message prints the exact versions):

[DIAMOND] Sensitive protein alignments at tree-of-life scale using DIAMOND.
          Buchfink B, Reuter K, Drost HG. 2021.
          Nature Methods 18, 366–368. https://doi.org/10.1038/s41592-021-01101-x

[MMSEQS2] MMseqs2 enables sensitive protein sequence searching for the analysis
          of massive data sets. Steinegger M & Söding J. 2017.
          Nat. Biotech. 35, 1026–1028. https://doi.org/10.1038/nbt.3988

[PRODIGAL] Prodigal: prokaryotic gene recognition and translation initiation
           site identification. Hyatt et al. 2010.
           BMC Bioinformatics 11, 119. https://doi.org/10.1186/1471-2105-11-119

Legacy v2 (eggNOG 5)

For eggNOG 5 databases, use the v2 branch or the last v2 release:

pip install eggnog-mapper==2.1.15

v2 and v3 databases are not interchangeable — v3 only works with eggNOG 7.