siRNA-Features

This repository contains the code and data accompanying the paper "siRNA Features - Reproducible Structure-Based Chemical Features for Off-Target Prediction".

Overview

This project introduces a framework for generating reproducible structure-based chemical features for siRNA, incorporating both molecular fingerprints and computationally derived siRNA-AGO2 complex structures. The framework enables better prediction of off-target effects in siRNA therapeutics through advanced feature engineering and machine learning approaches.

Key Features

• Structural prediction and modeling of siRNA-AGO2 complexes
• Extended Connectivity Fingerprint (ECFP) generation for siRNA modifications
• Multiple feature representation strategies (9 distinct datasets)
• Machine learning pipeline using AutoGluon
• Reproducible molecular dynamics simulations

Installation

# Clone the repository
git clone https://github.com/username/siRNA-Features.git
cd siRNA-Features

# Create and activate a conda environment
conda create -n sirna_features python=3.8
conda activate sirna_features

# Install dependencies
pip install -r requirements.txt

Project Structure

siRNA-Features/
├── data/                    # Dataset storage
│   ├── raw/                # Raw RNA-Seq data
│   └── processed/          # Processed features and results
├── docs/                   # Documentation
├── models/                 # Trained models
├── notebooks/             # Jupyter notebooks
│   ├── RNASeq/            # RNA-Seq analysis notebooks
│   ├── alignments/        # Structure alignment notebooks
│   └── modifications/     # Chemical modification notebooks
├── src/                   # Source code
│   ├── features/          # Feature generation
│   ├── models/           # Model training and evaluation
│   └── utils/            # Utility functions
├── tests/                # Unit tests
├── LICENSE
└── README.md

Usage

Feature Generation

1. Structural Prediction:

from src.features import structural_prediction
model = structural_prediction.ChaiDiscovery()
structure = model.predict_structure(sequence)

2. ECFP Generation:

from src.features import fingerprints
ecfps = fingerprints.generate_ecfp(molecule, nbits=512, radius=2)

3. Dataset Creation:

from src.features import dataset_builder
dataset = dataset_builder.create_dataset(structure, ecfps, type="dataset3")

Model Training

from src.models import training
model = training.AutoGluonTrainer(dataset)
model.train()

Datasets

The datasets used in this study are available at:

• RNA-Seq dataset: [Dataset Link]
• Processed features: [Features Link]
• Pre-trained models: [Models Link]

Results

Our framework achieved:

• AUPRC scores up to 0.784 (Dataset 3)
• Robust performance across multiple feature representations
• Reproducible structural predictions

Citation

If you use this code or find our work helpful, please cite:

@article{richter2024sirna,
  title={siRNA Features - Reproducible Structure-Based Chemical Features for Off-Target Prediction},
  author={Richter, Michael and Admasub, Alem},
  journal={Journal Name},
  year={2024}
}

License

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

Contributing

We welcome contributions! Please see our contributing guidelines for details.

Contact

• Michael Richter - [email protected]
• Alem Admasub - [email protected]