annorefine 2026.2.22

Genome annotation refinement using RNA-seq data

AnnoRefine

High-performance genome annotation refinement toolkit using RNA-seq data

AnnoRefine is a Rust-based toolkit for refining genome annotations and generating gene prediction hints from RNA-seq evidence. It provides both command-line tools and Python bindings for seamless integration into bioinformatics pipelines.

📖 Full Documentation

Features

• 🔧 UTR Refinement - Extend and trim UTRs based on RNA-seq coverage
• 🔀 Splice Site Refinement - Adjust intron boundaries using junction evidence
• 🆕 Novel Gene Detection - Discover new genes from RNA-seq data
• 🎯 Hint Generation - Convert BAM alignments to Augustus/GeneMark hints
• 📊 Hint Processing - Join and filter hints from multiple sources
• ⚡ High Performance - Multi-threaded Rust implementation
• 🐍 Python Bindings - Easy integration into Python workflows
• 🧭 Strand-Aware - Supports all RNA-seq library types (FR, RF, UU)

Installation

Python Package (Recommended):

pip install annorefine

Standalone Binary: Download from GitHub Releases

Build from Source:

git clone https://github.com/nextgenusfs/annorefine.git
cd annorefine
cargo build --release

See the Installation Guide for detailed instructions.

Quick Start

Python API:

import annorefine

# Refine annotations
result = annorefine.refine(
    fasta_file="genome.fa",
    gff3_file="annotations.gff3",
    bam_file="alignments.bam",
    output_file="refined.gff3"
)

# Generate hints for gene prediction
result = annorefine.bam2hints(
    bam_file="alignments.bam",
    output_file="hints.gff",
    library_type="RF",
    contig_map={'NC_000001.11': 'chr1'}  # Optional: rename contigs
)

# Join hints from multiple sources
result = annorefine.join_hints(
    input_files=["bam_hints.gff", "protein_hints.gff"],
    output_file="joined_hints.gff"
)

Command Line:

# Refine annotations
annorefine utrs \
    --fasta genome.fa \
    --gff3 annotations.gff3 \
    --bam alignments.bam \
    --output refined.gff3

# Generate hints
annorefine bam2hints \
    --in alignments.bam \
    --out hints.gff \
    --stranded RF

# Join hints
annorefine join-hints \
    --input bam_hints.gff protein_hints.gff \
    --output joined_hints.gff

See the User Guide for more examples.

Use Cases

• Annotation Refinement - Improve existing gene models with RNA-seq evidence
• Augustus Gene Prediction - Generate hints for ab initio gene prediction
• GeneMark-ETP - Create intron-only hints for GeneMark
• funannotate2 Integration - Seamless integration with gene prediction pipelines

Documentation

• 📖 User Guide
• 🐍 Python API Reference
• 💻 Command Line Reference
• 🚀 funannotate2 Integration

Performance

• Multi-threaded - Parallel processing with Rust backend
• Memory efficient - Streaming BAM processing
• Scalable - Handles mammalian-sized genomes efficiently

Typical performance:

• Human genome (~20K genes): 10-30 minutes on 8 cores
• Memory usage: 2-8 GB depending on genome size

Support

• 📖 Documentation
• 🐛 Bug Reports
• 💬 Discussions

Citation

Palmer, J. (2025). AnnoRefine: High-performance genome annotation refinement using RNA-seq data.
GitHub: https://github.com/nextgenusfs/annorefine

License

MIT License - see LICENSE file for details.

---

Built with ❤️ in Rust | Documentation | PyPI