Simple Neural Network

Simple Neural Network Model and ChemBERTaNeuralNetwork

Contains the SimpleNeuralNetwork and the ChemBERTaNeuralNetwork model.

class drevalpy.models.SimpleNeuralNetwork.simple_neural_network.ChemBERTaNeuralNetwork

Bases: SimpleNeuralNetwork

ChemBERTa Neural Network model using gene expression and ChemBERTa drug embeddings.

drug_views = ['chemberta_embeddings']
classmethod get_model_name()

Returns the model name.

Return type:

str

Returns:

ChemBERTaNeuralNetwork

load_drug_features(data_path, dataset_name)

Loads the ChemBERTa embeddings.

Parameters:

  • data_path (str) – Path to the ChemBERTa embeddings, e.g., data/

  • dataset_name (str) – name of the dataset, e.g., GDSC1

Return type:

FeatureDataset

Returns:

FeatureDataset containing the ChemBERTa embeddings

Raises:

FileNotFoundError – if the ChemBERTa embeddings file is not found

class drevalpy.models.SimpleNeuralNetwork.simple_neural_network.SimpleNeuralNetwork

Bases: DRPModel

Simple Feedforward Neural Network model with dropout using only gene expression data.

build_model(hyperparameters)

Builds the model from hyperparameters.

Parameters:

hyperparameters (dict) – includes units_per_layer and dropout_prob.

cell_line_views = ['gene_expression']
drug_views = ['fingerprints']
early_stopping = True
classmethod get_model_name()

Returns the model name.

Return type:

str

Returns:

SimpleNeuralNetwork

classmethod load(directory)

Load a trained SimpleNeuralNetwork instance from disk.

This includes: - model.pt: PyTorch state_dict of the trained model - hyperparameters.json: Dictionary with model hyperparameters - scaler.pkl: Fitted StandardScaler for gene expression features

Parameters:

directory (str) – Directory containing the saved model files

Return type:

SimpleNeuralNetwork

Returns:

An instance of SimpleNeuralNetwork with restored state

Raises:

FileNotFoundError – if any required file is missing

load_cell_line_features(data_path, dataset_name)

Loads the cell line features.

Parameters:

  • data_path (str) – Path to the gene expression and landmark genes

  • dataset_name (str) – name of the dataset

Return type:

FeatureDataset

Returns:

FeatureDataset containing the cell line gene expression features, filtered through the landmark genes

load_drug_features(data_path, dataset_name)

Loads the fingerprint data.

Parameters:

  • data_path (str) – Path to the fingerprints, e.g., data/

  • dataset_name (str) – name of the dataset, e.g., GDSC1

Return type:

FeatureDataset

Returns:

FeatureDataset containing the fingerprints

predict(cell_line_ids, drug_ids, cell_line_input, drug_input=None)

Predicts the response for the given input.

Parameters:

  • cell_line_ids (ndarray) – IDs of the cell lines to be predicted

  • drug_ids (ndarray) – IDs of the drugs to be predicted

  • cell_line_input (FeatureDataset) – gene expression of the test data

  • drug_input (FeatureDataset | None) – fingerprints of the test data

Return type:

ndarray

Returns:

the predicted drug responses

save(directory)

Save the trained model, hyperparameters, and gene expression scaler to the given directory.

This enables full reconstruction of the model using load.

Files saved: - model.pt: PyTorch state_dict of the trained model - hyperparameters.json: Dictionary containing all relevant model hyperparameters - scaler.pkl: Fitted StandardScaler for gene expression features

Parameters:

directory (str) – Target directory to store all model artifacts

Return type:

None

train(output, cell_line_input, drug_input=None, output_earlystopping=None, model_checkpoint_dir='checkpoints')

First scales the gene expression data and trains the model.

The gene expression data is first arcsinh transformed. Afterward, the StandardScaler() is fitted on the training gene expression data only. Then, it transforms all gene expression data. :type output: DrugResponseDataset :param output: training data associated with the response output :type cell_line_input: FeatureDataset :param cell_line_input: cell line omics features :type drug_input: FeatureDataset | None :param drug_input: drug omics features :type output_earlystopping: DrugResponseDataset | None :param output_earlystopping: optional early stopping dataset :type model_checkpoint_dir: str :param model_checkpoint_dir: directory to save the model checkpoints :raises ValueError: if drug_input (fingerprints) is missing

Return type:

None

Parameters:

Multi-OMICS Neural Network

Contains the baseline MultiOmicsNeuralNetwork model.

class drevalpy.models.SimpleNeuralNetwork.multiomics_neural_network.MultiOmicsNeuralNetwork

Bases: DRPModel

Simple Feedforward Neural Network model with dropout using multiple omics data.

build_model(hyperparameters)

Builds the model from hyperparameters.

The model is a simple feedforward neural network with dropout. The PCA is used to reduce the dimensionality of the methylation data.

Parameters:

hyperparameters (dict) – dictionary containing the hyperparameters units_per_layer, dropout_prob, and methylation_pca_components.

cell_line_views = ['gene_expression', 'methylation', 'mutations', 'copy_number_variation_gistic']
drug_views = ['fingerprints']
early_stopping = True
classmethod get_model_name()

Returns the model name.

Return type:

str

Returns:

MultiOmicsNeuralNetwork

classmethod load(directory)

Load a trained MultiOmicsNeuralNetwork instance from disk.

Required files: - model.pt - hyperparameters.json - gene_scaler.pkl - methylation_scaler.pkl - methylation_pca.pkl - metadata.json

Parameters:

directory (str) – Directory containing the saved model files

Return type:

MultiOmicsNeuralNetwork

Returns:

Fully restored MultiOmicsNeuralNetwork instance

Raises:

FileNotFoundError – if any required file is missing

load_cell_line_features(data_path, dataset_name)

Loads the cell line features.

Parameters:

  • data_path (str) – data path e.g. data/

  • dataset_name (str) – dataset name e.g. GDSC1

Return type:

FeatureDataset

Returns:

FeatureDataset containing the cell line omics features, filtered through the drug target genes

load_drug_features(data_path, dataset_name)

Load the drug features.

Parameters:

  • data_path (str) – path to the drug features, in this case the drug fingerprints, e.g., data/

  • dataset_name (str) – name of the dataset, e.g., GDSC1

Return type:

FeatureDataset

Returns:

FeatureDataset containing the drug fingerprint features

predict(cell_line_ids, drug_ids, cell_line_input, drug_input=None)

Applies arcsinh + scaling to gene expression and scaling + PCA to methylation, then predicts.

Parameters:
Return type:

ndarray

Returns:

predicted response

save(directory)

Save the trained model, hyperparameters, scalers, PCA object, and feature dimensions to disk.

Files saved: - model.pt - hyperparameters.json - gene_scaler.pkl - methylation_scaler.pkl - methylation_pca.pkl - metadata.json

Parameters:

directory (str) – Target directory

Return type:

None

train(output, cell_line_input, drug_input=None, output_earlystopping=None, model_checkpoint_dir='')

Fits the PCA and trains the model.

Parameters:
Raises:

ValueError – if drug_input (fingerprints) is missing

Model utils

Utility functions for the simple neural network models.

class drevalpy.models.SimpleNeuralNetwork.utils.FeedForwardNetwork(hyperparameters, input_dim)

Bases: LightningModule

Feed forward neural network for regression tasks with basic architecture.

Parameters:
configure_optimizers()

Overwrites the configure_optimizers from the LightningModule.

Return type:

Optimizer

Returns:

Adam optimizer

fit(output_train, cell_line_input, drug_input, cell_line_views, drug_views, output_earlystopping=None, trainer_params=None, batch_size=32, patience=5, num_workers=2, model_checkpoint_dir='checkpoints')

Fits the model.

First, the data is loaded using a DataLoader. Then, the model is trained using the Lightning Trainer. :type output_train: DrugResponseDataset :param output_train: Response values for training :type cell_line_input: FeatureDataset :param cell_line_input: Cell line features :type drug_input: FeatureDataset | None :param drug_input: Drug features :type cell_line_views: list[str] :param cell_line_views: Cell line info needed for this model :type drug_views: list[str] :param drug_views: Drug info needed for this model :type output_earlystopping: DrugResponseDataset | None :param output_earlystopping: Response values for early stopping :type trainer_params: dict | None :param trainer_params: custom parameters for the trainer :type batch_size: :param batch_size: batch size for the DataLoader, default is 32 :type patience: :param patience: patience for early stopping, default is 5 :type num_workers: int :param num_workers: number of workers for the DataLoader, default is 2 :type model_checkpoint_dir: str :param model_checkpoint_dir: directory to save the model checkpoints :raises ValueError: if drug_input is missing

Return type:

None

Parameters:
forward(x)

Forward pass of the model.

Parameters:

x – input data

Return type:

Tensor

Returns:

predicted response

predict(x)

Predicts the response for the given input.

Parameters:

x (ndarray) – input data

Return type:

ndarray

Returns:

predicted response

training_step(batch)

Overwrites the training step from the LightningModule.

Does a forward pass, calculates the loss and logs the loss. :type batch: :param batch: batch of data :returns: loss

validation_step(batch)

Overwrites the validation step from the LightningModule.

Does a forward pass, calculates the loss and logs the loss. :type batch: :param batch: batch of data :returns: loss

class drevalpy.models.SimpleNeuralNetwork.utils.RegressionDataset(output, cell_line_input, drug_input, cell_line_views, drug_views)

Bases: Dataset

Dataset for regression tasks for the data loader.

Parameters: