- The notebook EDA contains the code to process the raw drugs, cell lines gene expression, and mutation data for downstream procedures
- The TrainJTNN notebook in FastJTNNpy3 folder contains the code to train the Junction Tree Variational Autoencoder and obtain the latent representation for the drugs
- The Train_GeneAE notebook contains the code to train the autoencoder and obtain the latent representation of the cell lines
- The MLP notebook contains the code to train the model
- The Inference notebook contains the code to evaluate the model and run the heursitic algorith to identify candidate drugs