Home

Efesto Documentation

Efesto is a profile HMM annotation pipeline for iron-cycling and metal resistance genes in genomes and metagenomes. It extends FeGenie with coordinate-aware operon clustering, a cluster confidence scoring system, an expanded curated HMM library, and integration with external tools (UniOP, antiSMASH, Anvi'o).

---

Contents

Getting started

• Installation
• Usage — quick start

HMM library

• HMM library overview — categories, sources, model counts, registry format
• Iron-sulfur assembly models — SUF / ISC biology and HMM details
• Library curation log — deduplication decisions, deprecation rationale
• Biological rationale for new models — all new categories

Pipeline logic

• Operon rules — rule engine, JSON schema, TonB/ExbBD energizer-guard
• Cluster confidence scoring — formula and all four components
• AntiSMASH BGC integration — siderophore BGC boost

Outputs and integration

• Output formats — all files, all columns
• Anvi'o integration — functions import, gene-scores import
• Visualisation — R heatmap script

Development

• Extending the HMM library — adding models, curation workflow
• Publication plan — manuscript roadmap

---

Quick orientation

Input: FAA/FNA + GFF  →  hmmsearch  →  genomic clustering  →  operon rules
  →  confidence scoring  →  GFF3 + TSV + summary stats + optional Anvi'o TSVs

Key design decisions:

• One hit per ORF (highest bitscore wins across all HMMs).
• Operon clustering uses bp coordinates (GFF mode) or Prodigal ordinal index (fallback).
• Confidence score is multiplicative: hmm_weight × co_occ_score × uniop_pair_score × bgc_boost.
• TonB-ExbBD are energy-transducer proteins, not substrate-specific; a cluster of only Ton-motor proteins is dropped by the SIDERO_TRANSPORT rule (see Operon rules).
• Zero-cutoff models (MetHMMDB + 14 FeGenie siderophore models) use --zero_cutoff_min_bitscore as a bitscore floor instead of a calibrated threshold.

---

Citing Efesto

If you use Efesto, cite all sources that apply to your run. See README — Citations.