Genome Annotation Pipeline - Usage Guide

Overview

Complete genome annotation workflow from assembled contigs to functional annotation. Supports both prokaryotic (Bakta one-step) and eukaryotic (RepeatMasker, BRAKER3, eggNOG-mapper, Infernal) paths with QC checkpoints at each stage.

Prerequisites

# Prokaryotic path
conda install -c bioconda bakta
bakta_db download --output /path/to/bakta_db --type full

# Eukaryotic path
conda install -c bioconda repeatmodeler repeatmasker braker3 busco agat
conda install -c bioconda eggnog-mapper interproscan trnascan-se infernal

# Python dependencies
pip install pandas biopython

Input data: - Assembled genome FASTA (contigs or scaffolds) - For eukaryotes: RNA-seq BAM and/or protein evidence FASTA (e.g., OrthoDB)

Quick Start

Tell your AI agent what you want to do: - "Annotate my newly assembled bacterial genome" - "Run a full eukaryotic genome annotation pipeline" - "Annotate my fungal assembly with repeat masking and gene prediction" - "Set up a prokaryotic annotation with Bakta"

Example Prompts

Prokaryotic

"I have an assembled E. coli genome. Run Bakta annotation with NCBI-compatible locus tags."

"Annotate my bacterial contigs and check if the gene count looks reasonable."

Eukaryotic

"I have an assembled insect genome with RNA-seq alignments. Run the full annotation pipeline: repeats, gene prediction, functional annotation."

"Mask repeats in my plant genome, predict genes with BRAKER3, and annotate them with eggNOG-mapper."

QC Focus

"Check whether my genome annotation has reasonable BUSCO completeness and gene counts."

"My annotation only has 40% functional assignment. Help me troubleshoot."

What the Agent Will Do

1. Assess genome type (prokaryotic or eukaryotic)
2. For prokaryotes: run Bakta one-step annotation
3. For eukaryotes: mask repeats with RepeatModeler/RepeatMasker
4. Predict genes with BRAKER3 using available evidence
5. Run BUSCO to check annotation completeness
6. Assign functions with eggNOG-mapper and InterProScan
7. Annotate ncRNAs with tRNAscan-SE and Infernal
8. Merge annotations and validate final GFF3
9. Report QC metrics at each stage

Tips

• Always soft-mask (-xsmall) rather than hard-mask for gene prediction input
• BRAKER3 with both RNA-seq and protein evidence gives best results
• Use the full Bakta database for comprehensive prokaryotic annotation
• BUSCO completeness > 90% is the target; use the deepest applicable clade dataset (not the shallow eukaryota_odb10), and compare proteome-mode to genome-mode BUSCO on the same assembly to tell whether the predictor or the assembly is the limit (see genome-annotation/annotation-qc)
• If functional annotation is < 60%, try broader taxonomy scope in eggNOG
• Check repeat content against expectations for the taxon before proceeding
• Run QUAST and BUSCO on the assembly itself before starting annotation

Related Skills

• genome-annotation/prokaryotic-annotation - Bakta and Prokka details
• genome-annotation/eukaryotic-gene-prediction - BRAKER3 and AUGUSTUS options
• genome-annotation/repeat-annotation - Soft-masking before gene prediction
• genome-annotation/functional-annotation - eggNOG-mapper and InterProScan
• genome-annotation/ncrna-annotation - Infernal/Rfam and tRNAscan-SE detail
• rna-structure/ncrna-search - Covariance-model search, gathering thresholds, and clan resolution
• genome-annotation/annotation-qc - BUSCO genome-vs-proteome, OMArk, CheckM2 gates
• genome-assembly/assembly-qc - Pre-annotation assembly quality checks
• genome-intervals/gtf-gff-handling - GFF3/GTF hierarchy traversal, AGAT sanitizing/validation, coordinate conversion, and seqid-consistency checks on the merged annotation