The aim of this repo is to be able to predict the probability of secondary structure conformation based on the sequence of amino acid only with a Artifical intelligence model: ProtTrans. Once these probabilities computed, give personalised dyhedral angle potentials for each amino acid based on sequence information. This will make it possible to make proteins more flexible/regionally oriented and thus ensure better folding.
Before using this code, ensure that you have the following prerequisites in place:
- Python 3.7 or later
- PyTorch
- Transformers library
- H5py
- Numpy
- Pandas
Please find all information in requirements
If you plan to run this code locally, you need access to a GPU with at least 5.2GB of memory.
-
Clone the repository to your local machine.
git clone https://github.com/Romumrn/martini_AI cd martini_AI -
Install the required Python libraries using pip.
pip install -r requirements.txt
- Run the code for protein secondary structure prediction. You can provide a sequence as a command-line argument.
python run_martini_AI.py --sequence "SEQUENCE" --sequence_id "ID"
#or
python run_martini_AI.py --fasta protain.fasta
You can also try the test case provided on google collab
(Not tested yes)