ML-augmented-docking-CYP-inhibition

Code repository for Machine Learning-Augmented Docking. 1. CYP inhibition prediction. Authors: Benjamin Weiser, Jérôme Genzling , Mihai Burai Patrascu, Ophélie Rostain and Nicolas Moitessier

ML-augmented-docking-CYP-inhibition

Code repository for Machine Learning-Augmented Docking. 1. CYP inhibition prediction. Authors: Benjamin Weiser, Jérôme Genzling , Mihai Burai Patrascu, Ophélie Rostain and Nicolas Moitessier

Code Workflow Procedure

1. Cleaned and combined test and train Pei sets with CYP_clean_files.ipnyb. Sets combined and then clustered to create new train and test sets
2. Dock each ligand 5 times to its respective isoform using FITTED. Docked data can be found here: (to be inserted)
3. Create analogue sets using FITTED. Create max train-test similarity using CYP_TC_DataSets.py
4. Run RF with Feature Importances using max train-test similarity of 0.8 using ML_over_Tanimoto.py which calls CYP_inhibition_functions.py and Do_ML2.py
5. Using these selected features run all ML models on all datasets using ML_over_Tanimoto.py which calls CYP_inhibition_functions.py and Do_ML2.py
6. Use CYP_evaluate_and_ensemble.py which calls CYP_evaluate_and_ensemble_functions.py to make ensembles and evaluate and graph model performance

All code and folders focused in this project were in a folder called EnsembleLearning which you will have to change to your directory Can run code, CYP_evaluate_and_ensemble, to load and evaluate models from ../CSVFiles/TanimotoCSV_Final

Train Models

Trained model can be found in CSVFiles/TanimotoCSV_Final/Models MaxTC is model made using different max train-test Tanimoto shown in figure 5, 7, 8 MaxTCsamed is model made using different max train-test Tanimoto with same data size shown in figure 6 Analog is model trained using the anolog datasets shown in figure 9

Docking Data

Results_CSV_ZIPPED contains all docked data used To implement into the workflow for replication unzip into /EnsembleLearning/CSVFiles/TanimotoCSV_Final/Results_CSV

Datasets

In CSVFiles/TanimotoCSV_Final: 80TC_TESTSETS contains the test sets for MaxTC Analog_Clusters contains all the data for the analog sets Clusters_Max_TC contains all the MaxTC clusters Smiles contains all the original smiles of the Pei sets