Data and code for Reker et al Cell Rep 2020.
contains raw experimental data of the in vitro, ex vivo, in vivo validations.
contains raw chembl22 data for model training. Please note that this data is added for reproducibility, but we encourage users to curate updated versions of ChEMBL to ensure training models on larger and improved data. Cf. https://www.ebi.ac.uk/chembl/
Python code to run predictions of GRAS/IIG bioactivities use a respective target ID to train a specific model For example,
python prediction.py 11398
will run predictions for P-glycoprotein and store prediction results in a seperate TSV file
SMILES structures of GRAS/IIG compounds used in predictions
SMILES structures of molecular weight-matched background library of random chemicals
known IC50/Ki activities of GRAS/IIG compounds from ChEMBL22, serves as validation data for model optimization/selection