The implementation of the paper "Language models enable zero-shot prediction of RNA secondary structure including pseudoknots".

Install PyTorch 1.6+, python 3.7+

1. Clone the repo

git clone https://github.com/gongtiansu/RNA-km.git

2. Install python packages

cd RNA-km
pip install -r requirements.txt

3. Download pretrained model weight and place the pth file into the weight folder

mkdir weight
mv weight.pth weight 

1. RNA-km.py: extract RNA sequence representation (L * 1024) and attention maps from sequence (fasta format)

python RNA-km.py -i <RNA_fasta> -o <output_dictionary> (--attn) (--cuda)

cd example
./run_example.sh

Distributed under the MIT License. See LICENSE for more information.