This repository contains the data and script used for the manuscript "Machine learning for efficient prediction of protein redox potential: the flavoproteins case" submitted to the Journal of Chemical Information and Modeling.
The Flavoproteins Dataset is available at
data/dataset.xlsx. For each flavoprotein used in this work we report:
- PDB-ID
- Organism and Family
- Classification and Cofactor-type
- Technique resolution
- Em, E1, E2
- Reference of the experimental work
All the scripts used to reproduce the work are here reported:
- Em predict.py: for the ML pipeline used to test the performance of the various ML models;
- Features Calculator.py: features computation of the flavoprotein PDB files reported in data/dataset.xlsx;
- ML_models.py: contains all the machine learning models hparams;
- SHAP_analysis.ipynb: notebook for SHAP analysis;
- HeatMap.py: script to create the table of performance of the various estimator for different combinations of radii r1 and r2;
- env.yml: yaml file to create a mamba env from file.
We suggest to create a mamba env to run the script as follow:
- install mamba
- git clone https://github.com/CompBtBs/Prediction_Flavoprotein_EM.git
- cd Prediction_Flavoprotein_EM
- mamba env create -f env.yml
- install PyBioMed using pip
