TorRNA

Improved prediction of Torsion angles of RNA by leveraging large language models.

Setting up dependencies

The script slurm_setup_env.sh contains the commands to set up a minimal Anaconda environment on a HPC with a Slurm workload manager, the CUDA 11.6 toolkit can be loaded with the module load... command that is commented out in the script. The script can be run as follows:

source slurm_setup_env.sh

Data

The ./data folder has all the data required to train and test TorRNA.

Training TorRNA

main.py trains a model of TorRNA and saves model checkpoints in the ./checkpoints folder. We have provided the best TorRNA models in this folder.

Testing TorRNA

compare_predictions.py tests TorRNA on the test set, and also gives the prediction errors for SPOT-RNA-1D and a random baseline method.

Jupyter Notebooks

All the results in the manuscript were generated using the .ipynb files provided in this repository.