installation

Installation

---

Requirements

Dependency	Version	Role
HMMER	≥ 3.3	Profile HMM search
Prodigal / Pyrodigal	any	ORF calling (--fna_dir)
Python	≥ 3.9	Pipeline runtime
samtools	≥ 1.10	Coverage (--bam / --bams)
UniOP	any	Operon prediction (--operon_prediction)
eggNOG-mapper	2.x	Tier-2 confirmation for needs_confirmation-flagged hits (--run_eggnog; not needed for --eggnog_annotations) — optional, see below
R + pheatmap + plotly	any	Heatmap visualisation (optional)

---

Option 1 — Conda / Mamba (recommended)

git clone https://github.com/l-gallucci/Efesto.git
cd Efesto
mamba env create -f environment.yml
conda activate efesto
efesto --help

environment.yml installs all Python and non-Python dependencies and registers the efesto command via pip install -e ..

---

Option 2 — Add to an existing conda environment

conda activate YOUR_ENV
pip install -e /path/to/Efesto
# Then install HMMER and samtools separately if not already present
conda install -c bioconda hmmer samtools

---

Option 3 — Manual (no conda)

# Install HMMER system-wide or via your package manager
# Then:
pip install -e /path/to/Efesto

Ensure hmmsearch, hmmpress, and (if using --fna_dir) prodigal or pyrodigal-gv are on $PATH.

---

Installing UniOP (optional)

UniOP is required only for --operon_prediction. It has no conda package; install from source:

git clone https://github.com/hongsua/UniOP.git
# No build needed — pure Python + dependencies (numpy, scipy)
pip install numpy scipy

Pass the path to the UniOP script with --uniop_path /path/to/UniOP/src/UniOP.

---

Installing eggNOG-mapper (optional)

eggNOG-mapper is only needed for --run_eggnog (Efesto invoking it internally on the small needs_confirmation-flagged ORF subset). If you already have an .emapper.annotations file from a run you did elsewhere, use --eggnog_annotations /path/to/file instead — that path does not require eggNOG-mapper installed at all.

Install it into the same efesto environment. environment.yml caps python<3.12 specifically so this works — eggNOG-mapper's conda package also pins python<3.12, and since the efesto environment never exceeds that ceiling to begin with, adding eggNOG-mapper later never requires a Python downgrade. Verified directly: a fresh solve of Efesto's full dependency set together with eggnog-mapper resolves cleanly (conda picks Python 3.11.x), and the test suite passes running under that exact Python version.

conda activate efesto
conda install -c bioconda -c conda-forge eggnog-mapper
download_eggnog_data.py --data_dir /path/to/eggnog_db   # tens of GB, one-time
efesto ... --run_eggnog --eggnog_db_dir /path/to/eggnog_db

---

Verifying the installation

# Check CLI is available
efesto --help

# Verify HMM library integrity
python scripts/curate_hmm_library.py --verify hmm_library/

# Run tests
pytest tests/ -v

---

HMM library first run

On first use with a library that contains pre-HMMER3/f profiles, run:

efesto --normalize_hmms --faa_dir orfs/ --hmm_dir hmm_library/ --out results/

--normalize_hmms converts any legacy profiles in-place using hmmconvert before the search. It is safe to run repeatedly (already-current files are skipped). After the first conversion the flag is no longer needed.

---

R packages (optional, for heatmaps)

conda install -c conda-forge r-pheatmap r-plotly r-htmlwidgets r-optparse r-rcolorbrewer r-scales