Deep learning and optimization code to generate plastic binding peptides (PBPs)
camsol_calculation/
contains a script to compute the CamSol value for a given peptidedata/
contains PepBD datasets for PE- and PS-binding that are used to train the deep learning models as well as a small sample dataset with 500 peptides designed by MCTS and their predicted PepBD scores for demonstration purposeexamples/
contains jupyter notebooks to demonstrate how to run our model to generate peptides and how to perform SHAP analysis based on the trained score predictorspeptide_generators/
contains the MCTS generators described in the paperscore_predictors/
contains both the trained LSTM models for PE and PS binding prediction
- python 3.8.2
- pandas 1.4.1
- shap 0.41.0
- scikit-learn 1.2.2
- numpy 1.20.1
- tensorflow 2.6.0
TBD