MultiStructRNA

MultiStructRNA is a unified Python interface for RNA secondary structure prediction using multiple state-of-the-art algorithms. It provides a consistent API for predicting RNA secondary structures, calculating thermodynamic properties, and visualizing results.

Features

• Unified API: Single interface for multiple prediction algorithms
• Multiple Algorithms: Support for LinearFold, LinearPartition, ViennaRNA, and RNAstructure
• SHAPE Support: Incorporate experimental SHAPE reactivity data
• Visualization: Generate publication-ready structure plots
• Metrics: Calculate thermodynamic properties and structural metrics
• CLI: Command-line interface for quick predictions

Supported Algorithms

Algorithm	Type	Description
linearfold	MFE	Fast linear-time MFE structure prediction
linearpartition	Partition	Fast linear-time partition function calculation
viennafold	MFE	ViennaRNA MFE structure prediction
viennapartition	Partition	ViennaRNA partition function calculation
rnastructurefold	MFE	RNAstructure MFE prediction
rnastructurepartition	Partition	RNAstructure partition function

Installation

Using pixi (recommended)

pixi install
pixi run setup-deps  # Install external binaries

Using pip

pip install multistructrna
multistructrna setup  # Install external binaries

Using uv

uv pip install multistructrna

From source

git clone https://github.com/modernatx/multistructrna.git
cd multistructrna
pip install -e ".[dev]"

External Dependencies

Some algorithms require external binaries. Run the setup command to install them:

multistructrna setup

This will install:

• LinearFold/LinearPartition: Cloned and compiled from GitHub
• ViennaRNA: Check installation (install via conda: conda install -c bioconda viennarna)
• RNAstructure: Instructions for manual installation

MXFold2 is optional. For the most stable baseline setup:

multistructrna setup --no-mxfold2

If you need MXFold2 later, follow the dedicated environment instructions in docs/howto/installation.md.

Quick Start

Python API

from multistructrna import MSRNA

# Create an RNA object
rna = MSRNA(seq="GUCACUAGUCCCAUUUCCUUGGCCUGAAAUGGUUUGUGAGGAGUCCGCGU", name="my_sequence")

# Predict structure using LinearPartition
rna.predict(algorithm="linearpartition", output_dir="/tmp")

# Access results
print(f"Structure: {rna.dotbracket}")
print(f"Free Energy of Ensemble: {rna.fee} kcal/mol")
print(f"Average Unpaired Probability: {rna.aup:.4f}")
print(f"Maximum Ladder Distance: {rna.mld}")

# Visualize the structure
rna.plot(plot_type="secondary_structure", output_dir="./output", show=True)

Using SHAPE Data

# Predict with SHAPE reactivity constraints
rna.predict(
    algorithm="linearpartition",
    output_dir="/tmp",
    shape="/path/to/shape_file.shape",
)

Command Line

# Predict structure
multistructrna predict AUGCUAGCUAGCUAGC --algorithm linearfold

# Predict with options
multistructrna predict AUGCUAGCUAGCUAGC \
    --algorithm linearpartition \
    --name my_sequence \
    --output-dir ./results \
    --temperature 37.0

# Show version
multistructrna version

Output Metrics

Depending on the algorithm, MultiStructRNA provides:

Metric	Description	Algorithms
dotbracket	Secondary structure in dot-bracket notation	All
mfe	Minimum Free Energy (kcal/mol)	Fold algorithms
fee	Free Energy of Ensemble (kcal/mol)	Partition algorithms
aup	Average Unpaired Probability	Partition algorithms
mld	Maximum Ladder Distance	All
bpps_matrix	Base Pair Probability Matrix	Partition algorithms
avg_shannon_entropy	Average Shannon Entropy	Partition algorithms

Documentation

For full documentation, visit the docs directory or run:

mkdocs serve

Development

Setup Development Environment

# Using pixi
pixi install -e dev
pixi run test

# Using pip
pip install -e ".[dev]"
pytest tests/

Running Tests

pytest tests/ -v

Code Quality

# Format code
black multistructrna tests

# Lint code
ruff check multistructrna tests

# Type checking
mypy multistructrna

Citation

If you use MultiStructRNA in your research, please cite:

@software{multistructrna,
  title = {MultiStructRNA: A Unified Interface for RNA Secondary Structure Prediction},
  author = {Jadeja, Yashrajsinh},
  year = {2024},
  url = {https://github.com/modernatx/multistructrna}
}

Please also cite the underlying tools you use:

• LinearFold/LinearPartition: Huang et al., 2019
• ViennaRNA: Lorenz et al., 2011
• RNAstructure: Mathews Lab

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.

Acknowledgments

• LinearFold/LinearPartition: Liang Huang Lab
• ViennaRNA: TBI Vienna
• RNAstructure: Mathews Lab, University of Rochester
• RiboGraphViz: Das Lab, Stanford
• RNAvigate: Weeks Lab, UNC Chapel Hill