SkillAgentSearch skills...

bakta-genome-annotation

Annotate bacterial and archaeal genomes and plasmids with Bakta's Prodigal/HMM/diamond pipeline. Identifies CDS, ncRNA, tRNA, rRNA, tmRNA, sORFs, CRISPR arrays, oriC/oriV/oriT, and gaps against a curated UniRef-derived database.

Install / Use

npx skills add jaechang-hits/SciAgent-Skills --skill bakta-genome-annotation

Installs into whichever agent you are using.

About this skill
📄

SKILL.md

Installable skill definition

Quality Score

91/100

Category

Automation

Supported Platforms

Zed

Our assessment of bakta-genome-annotation

bakta-genome-annotation scores 91/100 on our quality scale, 1093rd of 2,771 Automation skills we index (top 40%).

Its SKILL.md is 23 KB long, well organised into 66 sections with 17 code examples: a thorough specification that gives an agent plenty to work with.

It has 367 GitHub stars, a meaningful sign that others use it.

Substance
30/30
Structure
20/20
Description
15/15
Adoption
11/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated about 3 months ago, so bakta-genome-annotation is actively maintained.
  • Our last check on 2026-11-20 found the source still online.
  • No license is declared. By default that means all rights are reserved: you can read it, but reusing or redistributing it is not clearly permitted. Ask the author before building on it commercially.
  • Its trust signals score 88/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-10-05. Automated pattern scan on 2026-10-05. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

bakta-genome-annotation compared with similar skills

All 4 of these similar skills score higher than bakta-genome-annotation; compare them before choosing.

SkillScoreStarsUpdatedFormat
bakta-genome-annotation (this skill)by jaechang-hits913673mo agoSKILL.md
Agent-Reachby Panniantong100113.3k12d agoCLAUDE.md
rufloby ruvnet10082.1ktodayCLAUDE.md
headroomby headroomlabs-ai10080.2ktodayCLAUDE.md
Claude-Code-Game-Studiosby Donchitos10040.2k10d agoCLAUDE.md

Frequently asked questions

How do I install bakta-genome-annotation?
Run npx skills add jaechang-hits/SciAgent-Skills --skill bakta-genome-annotation. The install tabs above show the steps for each supported agent.
Which AI agents does bakta-genome-annotation work with?
It is written for Zed, as a SKILL.md file. Other agents that read the same format can often use it too.
Is bakta-genome-annotation safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It declares no license and scores 88/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is bakta-genome-annotation still maintained?
The repository was last updated about 3 months ago, so bakta-genome-annotation is actively maintained.

name: "bakta-genome-annotation" description: "Annotate bacterial and archaeal genomes and plasmids with Bakta's Prodigal/HMM/diamond pipeline. Identifies CDS, ncRNA, tRNA, rRNA, tmRNA, sORFs, CRISPR arrays, oriC/oriV/oriT, and gaps against a curated UniRef-derived database. Produces NCBI-compatible GFF3, GenBank, EMBL, JSON, FASTA, TSV, and a circular genome plot. Use Prokka for legacy pipelines or non-bacterial kingdoms; PGAP for NCBI GenBank submission." license: "GPL-3.0"

Bakta Genome Annotation

Overview

Bakta is a command-line pipeline for rapid, standardized annotation of bacterial and archaeal genomes and plasmids. It combines Prodigal for CDS prediction, tRNAscan-SE/Aragorn/Barrnap/Infernal for non-coding RNA, PILER-CR/PILERCR for CRISPR detection, and a tiered DIAMOND/HMM search against a curated UniRef100 + IPS/UPS database to assign gene names, EC numbers, GO terms, and COG categories. Bakta produces NCBI-compatible outputs (GFF3, GenBank, EMBL, INSDC-formatted FASTA, plus a JSON summary and a circular Circos plot) for a typical 5 Mb genome in 5–15 minutes on 8 CPUs.

When to Use

  • Annotating bacterial or archaeal genome assemblies (Illumina, PacBio, Nanopore) with NCBI-compatible locus tags and product names
  • Annotating plasmids and other circular replicons separately with --plasmid and --complete flags
  • Producing JSON-structured annotation outputs that can be parsed without GenBank or GFF3 detours
  • Generating a publication-ready circular genome plot via the bundled bakta_plot command
  • Annotating MAGs (metagenome-assembled genomes) with --meta to disable Prodigal training
  • Use Prokka instead when you need viral/mitochondrial kingdoms or when you must reproduce a legacy Prokka pipeline exactly
  • Use PGAP instead when submitting to NCBI GenBank with full standards compliance
  • Use Bakta when you want faster runs, regularly updated UniRef-derived databases, AMRFinderPlus integration, and a JSON summary out of the box

Prerequisites

  • Software: Bakta ≥ 1.9, Python 3.8+, Prodigal, tRNAscan-SE, Aragorn, Barrnap, Infernal, DIAMOND, HMMER3, PILER-CR, BLAST+, AMRFinderPlus
  • Database: Bakta DB (full ~70 GB, or light ~3 GB) downloaded once with bakta_db download
  • Python packages (for output parsing): biopython, pandas, matplotlib
  • Input: assembled genome in FASTA format (one or more contigs)
  • Hardware: ≥ 16 GB RAM for full DB, ≥ 4 GB RAM for light DB; ≥ 8 CPUs recommended

Check before installing: The tool may already be available in the current environment (e.g., inside a pixi / conda env). Run command -v bakta first and skip the install commands below if it returns a path. When running inside a pixi project, invoke the tool via pixi run bakta rather than bare bakta.

# Install Bakta via conda/mamba (recommended)
mamba install -c conda-forge -c bioconda bakta

# Verify installation
bakta --version
# bakta 1.9.4

# Download the light database (~3 GB, faster, fewer functional hits)
bakta_db download --output db/ --type light

# Or full database (~70 GB, comprehensive UniRef100 coverage)
# bakta_db download --output db/ --type full

# Install Python parsing dependencies
pip install biopython pandas matplotlib

Pre-flight Interview

Settle these with the user before writing any analysis code.

decisions:
  - id: D1
    param: database
    kind: required
    source: user
    ask: "Annotate against the light database or the full one? The light database leaves more genes described only generically."
    default: null

  - id: D2
    param: organismMetadata
    kind: required
    source: user
    ask: "Which genus, species, and strain should be recorded in the output?"
    default: null

  - id: D3
    param: repliconTopology
    kind: required
    source: data
    ask: "Are these contigs complete circular replicons, plasmids, or draft fragments?"
    default: "draft fragments - no circularity assumed"

  - id: D4
    param: assemblyOrigin
    kind: required
    source: data
    ask: "Is this an isolate genome, or a bin recovered from a metagenome?"
    default: "isolate - gene prediction trains on the assembly itself"

  - id: D5
    param: translationTable
    kind: required
    source: user
    ask: "Which genetic code does this organism use?"
    default: "11, the bacterial and archaeal code"

  - id: D6
    param: locusTagPrefix
    kind: optional
    source: user
    ask: "Which prefix should locus tags carry, if they will be submitted or cross-referenced?"
    default: "generated automatically"

  - id: D7
    param: minContigLength
    kind: optional_conditional
    source: data
    ask: "Should short contigs be skipped?"
    default: "no minimum"

  - id: D8
    param: threadsAndPlotting
    kind: never_ask
    source: data
    reason: "Thread count and the circular plot affect runtime and output files, not the annotation"
    default: "available cores, plot enabled"

D5 is asked rather than defaulted because the exceptions are real and silent: Mycoplasma and relatives reassign a stop codon, so annotating them with the standard bacterial code truncates a large fraction of their genes into fragments that still look like ordinary short CDS calls.

Quick Start

# Annotate a bacterial genome — results in results/ directory
bakta genome.fasta \
    --db db/bakta_db_light \
    --output results/ \
    --prefix sample1 \
    --threads 8

# Inspect the JSON summary for feature counts
python -c "
import json
with open('results/sample1.json') as f:
    d = json.load(f)
print('Genus:', d['genome'].get('genus'))
print('Length:', d['genome']['size'], 'bp')
print('CDS:', sum(1 for f in d['features'] if f['type'] == 'cds'))
print('tRNA:', sum(1 for f in d['features'] if f['type'] == 'tRNA'))
"

Workflow

Step 1: Install Bakta and Download the Database

Install Bakta and prepare the reference database. The database download is one-time and reused across runs.

# Create a dedicated conda environment (avoids dependency conflicts)
mamba create -n bakta_env -c conda-forge -c bioconda bakta python=3.11 -y
mamba activate bakta_env

# Verify Bakta and its dependencies
bakta --version
# bakta 1.9.4

bakta --help | head -20

# Download the light database (sufficient for routine annotation)
mkdir -p db/
bakta_db download --output db/ --type light
# Downloads ~3 GB; expands to ~5 GB on disk

# Verify the database was extracted correctly
ls db/bakta_db_light/
# antifam.h3f  bakta.db  expert  oric.fna  pfam.h3f  rfam-go.tsv  ...

# (Optional) Update AMRFinderPlus DB used by Bakta for AMR gene calling
amrfinder -u

# Install Python parsing tools
pip install biopython pandas matplotlib

Step 2: Prepare the Input Assembly

Bakta requires clean FASTA headers without spaces or special characters. Pre-clean and optionally filter short contigs.

from Bio import SeqIO
import re

input_fasta = "genome.fasta"
records = list(SeqIO.parse(input_fasta, "fasta"))
print(f"Input assembly: {len(records)} contigs")
total_bases = sum(len(r) for r in records)
print(f"Total bases: {total_bases:,}")
print(f"Largest contig: {max(len(r) for r in records):,} bp")

# Bakta preferred: short, alphanumeric, unique IDs
cleaned = []
for i, rec in enumerate(records, 1):
    new_id = f"contig_{i:04d}"
    new_rec = rec.__class__(rec.seq, id=new_id, description="")
    cleaned.append(new_rec)

SeqIO.write(cleaned, "genome_clean.fasta", "fasta")
print(f"Wrote genome_clean.fasta with {len(cleaned)} contigs")
# Filter out short contigs (<200 bp) which contribute little to annotation
awk 'BEGIN{RS=">"; ORS=""} NR>1 {n=split($0, a, "\n"); seq=""; for(i=2;i<=n;i++) seq=seq a[i]; if (length(seq) >= 200) print ">" $0}' \
    genome_clean.fasta > genome_filtered.fasta

echo "Filtered assembly: $(grep -c '>' genome_filtered.fasta) contigs"

Step 3: Run Standard Bakta Annotation

Run Bakta with genus/species hints. Locus tags are auto-generated from the strain field.

# Standard annotation for a draft bacterial genome
bakta genome_clean.fasta \
    --db db/bakta_db_light \
    --output annotation/ \
    --prefix E_coli_K12 \
    --genus Escherichia \
    --species coli \
    --strain K12 \
    --locus-tag ECOLI \
    --threads 8 \
    --keep-contig-headers

# Expected runtime: 5–15 min for ~5 Mb genome on 8 CPUs (light DB)

echo "Bakta annotation outputs:"
ls annotation/
# E_coli_K12.embl   E_coli_K12.faa     E_coli_K12.ffn
# E_coli_K12.fna    E_coli_K12.gbff    E_coli_K12.gff3
# E_coli_K12.hypotheticals.faa  E_coli_K12.hypotheticals.tsv
# E_coli_K12.json   E_coli_K12.log     E_coli_K12.png
# E_coli_K12.svg    E_coli_K12.tsv     E_coli_K12.txt

Step 4: Parse the JSON Summary

Bakta's JSON output is the canonical, machine-readable annotation. Parse it directly for downstream pipelines.

import json
import pandas as pd
from collections import Counter

with open("annotation/E_coli_K12.json") as f:
    bakta = json.load(f)

# Genome-level metadata
genome = bakta["genome"]
print(f"Organism: {genome.get('genus')} {genome.get('species')} {genome.get('strain')}")
print(f"Size: {genome['size']:,} bp across {len(bakta['sequences'])} sequences")
print(f"GC content: {genome['gc']:.2%}")

# Feature type counts
features = bakta["features"]
type_counts = Counter(f["type"] for f in features)
print("\nFeature counts:")
for ftype, n in sorted(type_counts.items(), key=lambda x: -x[1]):
    print(f"  {ftype:>10}: {n}")

# Build a tidy CDS DataFrame
cds_rows = []
for f in features:
    if f["type"] != "cds":
        continue
    cds_rows.append({
        "locus_tag": f.get("locus", ""),
        "contig":    f.get("contig", ""),
        "start":     f.get("start"),
        "stop":      f.get("stop"),
        "strand":    f.get("strand"),
        "gene":      f.get("gene", ""),
        "product":   f.get("product", ""),
        "length_aa": len(f.get("aa", "")),
    })

cds_df = pd.DataFrame(cds_rows)
print(f"\nTotal CDS: {len(cds_df)}")
print(cds_df.head(5).to_string(index=False))

Step 5: Parse the TSV Feature Table

The TSV output is convenient for spreadsheet workflows and quick filtering.

import pandas as pd

# Bakta TSV begins with comment lines starting with '#'
df = pd.read_csv("annotation/E_coli_K12.tsv", sep="\t", comment="#",
                 names=["sequence_id", "type", "start", "stop", "strand",
                        "locus_tag", "gene", "product", "dbxrefs"])
print(f"Total features: {len(df)}")
print(f"Feature types: {df['type'].value_counts().to_dict()}")

# Hypothetical vs annotated CDS
cds = df[df["type"] == "cds"].copy()
hypothetical = cds["product"].str.contains("hypothetical", case=False, na=True)
print(f"\nCDS with assigned function: {(~hypothetical).sum()} / {len(cds)}")
print(f"Hypothetical proteins: {hypothetical.sum()}")

# Cross-references (UniRef, KEGG, EC, GO, etc.) parsed from the dbxrefs column
def split_xrefs(xref_str):
    if not isinstance(xref_str, str) or xref_str in ("", "-"):
        return []
    return [x.strip() for x in xref_str.split(",")]

cds["dbxref_list"] = cds["dbxrefs"].apply(split_xrefs)
ec_hits = cds[cds["dbxref_list"].apply(lambda xs: any(x.startswith("EC:") for x in xs))]
print(f"CDS with EC numbers: {len(ec_hits)}")
print(ec_hits[["locus_tag", "gene", "product"]].head(5).to_string(index=False))

Step 6: Render the Circular Genome Plot

Bakta emits a Circos-style PNG/SVG by default. Regenerate with custom styling using bakta_plot.

# Re-render the plot from the existing JSON with a different style
bakta_plot --output annotation/ \
           --prefix E_coli_K12_replot \
           --type cog \
           --dpi 300 \
           annotation/E_coli_K12.json

ls annotation/E_coli_K12_replot.*
# E_coli_K12_replot.png  E_coli_K12_replot.svg

# Display the rendered PNG inline in a Jupyter notebook
python <<'PY'
from pathlib import Path
import matplotlib.pyplot as plt

Truncated for display — read the full file on GitHub.

Related Skills

View on GitHub
GitHub Stars367
CategoryAutomation
Updated2mo ago
Forks36

Languages

Python

Trust signals

88/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

1 medium