The following repo consists of 2 main tools + an auxiliary one, namely:
- TLStab
- TLImm
- TLBind
To use these tools, you will need to have either conda or mamba installed. We highly recommend mamba, as it handles dependencies much faster and more reliably. To install mamba follow the instructions here. If you prefer to use conda instead, you can find instructions here, if it's not already installed in your machine.
First, clone the repo, and get into the main folder of the repo (where this README file is found). When either conda or mamba are sucessfully installed, you can create an enviroment by simply running:
mamba env create -f TLenv.yml
or
conda env create -f TLenv.yml
in the folder, followed by
mamba activate TLenv
or
conda activate TLenv
in order to run the tools.
It seems that, when using conda on Windows, there is a conflict between the pytorch installation and with some of the other packages. This happened to us on Windows only. A workaround that worked for us was to remove the pytorch entry from the YAML environment file, and install pytorch later when into the environment, using a pytorch installation command taken from here. Please reach out to us if you have any issues installing the required packages.
TLStab is our transfer learning-based pMHC stability predictor. You will find the corresponding files and instructions on how to use in the TLStab folder.
TLImm is our transfer learning-based peptide immunogenicity predictor. You will find the corresponding files and instructions on how to use in the TLImm folder.
TLBind is our pretrained binding affinity/eluted ligand predictor. You will find the corresponding files and instructions on how to use in the TLBind folder.