Machine learning meets pK_a

Prerequisites

The Python dependencies are:

• Python >= 3.7
• NumPy >= 1.18
• Scikit-Learn >= 0.22
• RDKit >= 2019.09.3
• Pandas >= 0.25
• XGBoost >= 0.90
• JupyterLab >= 1.2
• Matplotlib >= 3.1
• Seaborn >= 0.9

For the data preparation pipeline, ChemAxon Marvin^[1] is required, to use the prediction model with the included Python script, ChemAxon Marvin^[1] is not required. By default OpenEye QUACPAC/Tautomers^[2] is used for tautomer and charge standardization. If you want to use RDKit^[3] instead, you can use the --no-openeye flag for the run_pipeline.sh script as well as for the train_model.py and predict_sdf.py scripts.

Of course, you also need the code from this repository folder.

Installing

First of all you need a working Miniconda/Anaconda installation. You can get Miniconda at https://conda.io/en/latest/miniconda.html.

Now you can create an environment named "ml_pka" with all needed dependencies and activate it with:

conda env create -f environment.yml
conda activate ml_pka

You can also create a new environment by yourself and install all dependencies without the environment.yml file:

conda create -n ml_pka -c conda-forge python=3.10
conda activate ml_pka
conda install -c conda-forge scikit-learn rdkit xgboost jupyterlab matplotlib seaborn

Usage

Preparation pipeline

To use the data preparation pipeline you have to be in the repository folder and your conda environment have to be activated. Additionally, the Marvin commandline tool cxcalc and, if you don't use the --no-openeye flag, the QUACPAC commandline tool tautomers have to be contained in your PATH variable.

Also, the environment variables OE_LICENSE (containing the path to your OpenEye license file) if used and JAVA_HOME (referring to the Java installation folder, which is needed for cxcalc) have to be set.

After preparation, you can display a small usage information with bash run_pipeline.sh -h. Example call:

bash run_pipeline.sh --train datasets/chembl25.sdf --test datasets/AvLiLuMoVe.sdf

Prediction tool

First of all you have to be in the repository folder and your conda environment have to be activated. To use the prediction tool you have to retrain the machine learning model. Therefore, just call the training script, it will train the 5-fold cross-validated Random Forest machine learning model using 12 cpu cores. If you want to adjust the number of cores you can use the parameter --num-processes. If you want to use the dataset that was prepared without the usage of QUACPAC/Tautomers you can use the --no-openeye flag. Example call:

python train_model.py

If you used QUACPAC/Tautomers for dataset preparation it has to be available to use the prediction tool as it was mentioned in the chapter above. If not, you have to use the --no-openeye flag for the prediction tool as well.

Now you can call the python script with a SDF file and an output path:

python predict_sdf.py my_test_file.sdf my_output_file.sdf

NOTE: This model was build for monoprotic structures regarding a pH range of 2 to 12. If the model is used with multiprotic structures, the predicted values will probably not be correct.

Datasets

1. AvLiLuMoVe.sdf - Manually combined literature pK_a data^[3]
2. chembl25.sdf - Experimental pK_a data extracted from ChEMBL25^[4]
3. datawarrior.sdf - pK_a data shipped with DataWarrior^[5]
4. combined_training_datasets_unique.sdf - Preprocessed and combined data from datasets (2) and (3), used as training dataset and prepared with QUACPAC/Tautomers^[2]
5. combined_training_datasets_unique_no_oe.sdf - Preprocessed and combined data from datasets (2) and (3), prepared with RDKit MolVS instead of QUACPAC/Tautomers^[2]
6. AvLiLuMoVe_cleaned_mono_unique_notraindata.sdf - Preprocessed data from dataset (1), used as external testset
7. novartis_cleaned_mono_unique_notraindata.sdf - Preprocessed data from an inhouse dataset provided by Novartis^[6], used as external testset

Authors

Marcel Baltruschat - GitHub, E-Mail
Paul Czodrowski - GitHub, E-Mail

License

This project is licensed under the MIT License - see the LICENSE.md file for details.

References

[1] Marvin 20.1.0, 2020, ChemAxon, http://www.chemaxon.com
[2] QUACPAC 2.0.2.2: OpenEye Scientific Software, Santa Fe, NM. http://www.eyesopen.com
[3] Settimo, L., Bellman, K. & Knegtel, R.M.A. Pharm Res (2014) 31: 1082. https://doi.org/10.1007/s11095-013-1232-z
[4] Gaulton A, Hersey A, Nowotka M, Bento AP, Chambers J, Mendez D, Mutowo P, Atkinson F, Bellis LJ, Cibrián-Uhalte E, Davies M, Dedman N, Karlsson A, Magariños MP, Overington JP, Papadatos G, Smit I, Leach AR. (2017) 'The ChEMBL database in 2017.' Nucleic Acids Res., 45(D1) D945-D954.
[5] Thomas Sander, Joel Freyss, Modest von Korff, Christian Rufener. DataWarrior: An Open-Source Program For Chemistry Aware Data Visualization And Analysis. J Chem Inf Model 2015, 55, 460-473, doi 10.1021/ci500588j
[6] Richard A. Lewis, Stephane Rodde, Novartis Pharma AG, Basel, Switzerland