Joyanne Ma, Akhila Raj, Dorra Tray

We predict rat oral LD50 from molecular SMILES strings using:

• Morgan fingerprints + physicochemical descriptors (RDKit)
• XGBoost + Random Forest ensemble
• SHAP analysis for interpretability

• Test R² = 0.6368
• Test MAE = 0.4241
• Test RMSE = 0.5697

1. Open in Google Colab
2. Run all cells top to bottom
3. Final test evaluation is in the last section

pyTDC, rdkit, xgboost, scikit-learn, shap, pandas, numpy, matplotlib, seaborn, pillow