Skip to content

Latest commit

 

History

History
8 lines (5 loc) · 476 Bytes

README.md

File metadata and controls

8 lines (5 loc) · 476 Bytes

antibiotic

These notebooks reproduce the work in "A Deep Learning Approach to Antibiotic Discovery"

01_halicin_train.ipynb - train a ChemProp model to predict antibiotic activity.
02_halicin_predict.ipynb - use the model above to generate predictions for the Broad Drug Repurposing Collection.
03_antibiotic_xgboost.ipynb - an alternate approach using fingerprints and XGBoost rather than MPNN.