TODO: COMPLETE Funannotate: A Tool for Genome Annotation

Setting Up Funannotate with Conda and Docker

Funannotate is an automated genome annotation pipeline primarily designed for fungi, but it can also handle higher eukaryotes. It orchestrates repeat masking, gene prediction, and functional annotation into a streamlined workflow.

This guide covers:

1. Prerequisites
2. Installing Funannotate via Conda (Bioconda)
3. Quickstart with Docker
4. Optional: Installing GeneMark-ES/ET
5. Example Workflows
   • Genome + RNA-seq annotation
   • Genome-only annotation
6. Tips & Best Practices
7. Further Resources

1. Prerequisites

• Conda (Miniconda or Anaconda) for package management
• Docker (optional) for containerized runs
• Genome assembly (FASTA) of your target organism
• RNA-seq FASTQ files (paired or single-end) for evidence-based annotation (optional)
• Familiarity with basic command-line operations

2. Installing Funannotate via Conda

The recommended install uses Bioconda channels. If dependency resolution is slow, consider replacing conda with mamba.

# Configure channels
conda config --add channels defaults
conda config --add channels bioconda
conda config --add channels conda-forge

# Create and activate environment
conda create -n funannotate_env "python>=3.6,<3.9" funannotate -y
conda activate funannotate_env

Tip: If solving is slow:

conda install -n base mamba -y
mamba create -n funannotate_env funannotate -y

To update:

conda update funannotate -y

3. Quickstart with Docker

A Docker image with Funannotate and required databases is available.

# Pull the full image (with databases)
docker pull nextgenusfs/funannotate

# (Optional) Download the provided bash wrapper
wget -O funannotate-docker \
  https://raw.githubusercontent.com/nextgenusfs/funannotate/master/funannotate-docker
chmod +x funannotate-docker

# Run a command, e.g. predict step with 12 CPUs
./funannotate-docker predict -t predict --cpus 12

If you prefer a slim image (no databases):

docker pull nextgenusfs/funannotate-slim

You can use docker run directly or via the funannotate-docker wrapper to automatically bind volumes and retain your user permissions.

4. Optional: Installing GeneMark-ES/ET

GeneMark is not distributable via Bioconda. To enable GeneMark support:

1. Register and download from GeneMark license page.
2. Extract and edit shebangs of Perl scripts to use /usr/bin/env perl.
3. Add the gmes_petap.pl directory to your $PATH, or set:

   export GENEMARK_PATH=/path/to/gmes_petap
   

Without GeneMark, Funannotate will rely on BUSCO/Augustus for ab initio predictions.

5. Example Workflows

5.1 Genome + RNA-seq Annotation

Assume files in working directory:

assembly.fasta  
left_R1.fq.gz   right_R1.fq.gz  
left_R2.fq.gz   right_R2.fq.gz  
nanopore_rna.fq.gz  

# 1) Clean and sort assembly
funannotate clean -i assembly.fasta --minlen 1000 -o assembly.clean.fa
funannotate sort -i assembly.clean.fa -b scaffold -o assembly.sorted.fa

# 2) Mask repeats
funannotate mask -i assembly.sorted.fa --cpus 12 -o assembly.masked.fa

# 3) Train with RNA-seq
funannotate train \
  -i assembly.masked.fa \
  --left left_R1.fq.gz,right_R1.fq.gz \
  --right left_R2.fq.gz,right_R2.fq.gz \
  --nanopore_mrna nanopore_rna.fq.gz \
  --stranded RF --jaccard_clip \
  --species "MySpecies" --strain "StrainA" \
  --cpus 12 -o fun_run

# 4) Predict gene models
funannotate predict -i assembly.masked.fa \
  --species "MySpecies" --strain "StrainA" \
  --cpus 12 -o fun_run

# 5) Update UTRs and refine gene models
funannotate update -i fun_run --cpus 12

# 6) Functional annotation
funannotate iprscan -i fun_run -m docker --cpus 12
funannotate annotate -i fun_run --cpus 12

Results appear under fun_run/predict_results, fun_run/update_results, and fun_run/annotate_results.

5.2 Genome-Only Annotation

Without RNA data, skip train/update and use BUSCO seed species:

# Mask repeats
funannotate mask -i assembly.fasta --cpus 12 -o assembly.masked.fa

# Predict with BUSCO-based training
funannotate predict \
  -i assembly.masked.fa \
  --species "MySpecies" --strain "StrainA" \
  --busco_seed_species botrytis_cinerea \
  --cpus 12 -o fun_run_genome_only

# Functional annotation
funannotate annotate -i fun_run_genome_only --cpus 12

6. Tips & Best Practices

• --dont_overwrite: protect existing outputs.
• Resume pipeline: use --resume if interrupted.
• Metagenomes: add --meta in predict step for uneven coverage.
• Max intron length: adjust with --max_intronlen for non-fungal genomes.
• Repeat-aware EVM: use --repeats2evm in predict to reduce false positives in large genomes.

7. Further Resources

• Funannotate docs: http://funannotate.readthedocs.io
• GitHub: https://github.com/nextgenusfs/funannotate
• Bandage for GFA visualization: https://github.com/rrwick/Bandage

Happy annotating!