Data and scripts required to generate the plots for:
Active learning driven prioritisation of compounds from on-demand libraries targeting the SARS-CoV-2 main protease, by Ben Cree, Mateusz Bieniek, Siddique Amin, Akane Kawamura and Daniel Cole.
Please see the preprint for further details.
Install the environment via:
mamba create --name fegrow_al_data python=3.11 jupyter rdkit numpy==1.24.4 scikit-learn umap-learn seaborn chemplot -c conda-forge -c anaconda
conda activate fegrow_al_data
pip install useful_rdkit_utils==0.2.7
To run the jupyter notebook:
python -m ipykernel install --user --name=fegrow_al_data
jupyter notebook
and open the notebook file plots.ipynb.
The file cs_49k.csv contains the SMILES and predicted affinity of the oracle dataset.
The folders rep_* contain five replicas of active learning hyperparameter tuning. For each experiment, folders contain SMILES and predicted affinity of compounds selected at each active learning cycle.
The same information is provided for the four prospective runs in the folders mpro-*, for the four different objective functions used.