QSAR Model for ChEMBL Targets

This app uses a neural network model to predict the pIC50 values for compounds against two ChEMBL targets:

• CHEMBL238 - Epidermal growth factor receptor erbB1
• CHEMBL228 - Cytochrome P450 19A1

The model was trained on IC50 data from ChEMBL, preprocessed, and optimized using Optuna.