feat: stable

DeepDRA.py

@@ -170,7 +170,7 @@ def train(model, train_loader, val_loader,  num_epochs,class_weights):
                 mlp_loss_val = mlp_weight * mlp_loss_fn(mlp_output_val, val_target)
 
                 # Total loss is the sum of autoencoder losses and MLP loss
-                total_val_loss = drug_ae_loss_val + cell_ae_loss_val + mlp_loss_val
+                total_val_loss = (drug_ae_loss_val + cell_ae_loss_val + mlp_loss_val).item()
 
                 # Calculate accuracy
                 val_predictions = torch.round(mlp_output_val)

data_loader.py

@@ -134,7 +134,6 @@ class RawDataLoader:
                 df.columns = df.columns.str.replace('_cell_mut', '')
                 df.columns = df.columns.str.replace('_cell_CN', '')
                 df.columns = df.columns.str.replace('_cell_exp', '')
-
                 df = (df - df.min()) / (df.max() - df.min())
                 df = df.fillna(0)
                 if intersect:

main.py

@@ -161,8 +161,8 @@ def run(k, is_test=False ):
 
     # Step 2: Load training data
     train_data, train_drug_screen = RawDataLoader.load_data(data_modalities=DATA_MODALITIES,
-                                                            raw_file_directory=GDSC_RAW_DATA_FOLDER,
-                                                            screen_file_directory=GDSC_SCREENING_DATA_FOLDER,
+                                                            raw_file_directory=RAW_BOTH_DATA_FOLDER,
+                                                            screen_file_directory=BOTH_SCREENING_DATA_FOLDER,
                                                             sep="\t")
 
 
@@ -170,10 +170,15 @@ def run(k, is_test=False ):
     if is_test:
         test_data, test_drug_screen = RawDataLoader.load_data(data_modalities=DATA_MODALITIES,
                                                               raw_file_directory=CCLE_RAW_DATA_FOLDER,
-                                                              screen_file_directory=CTRP_SCREENING_DATA_FOLDER,
+                                                              screen_file_directory=CCLE_SCREENING_DATA_FOLDER,
                                                               sep="\t")
         train_data, test_data = RawDataLoader.data_features_intersect(train_data, test_data)
 
+        # common_columns = list(set(train_drug_screen.columns) & set(test_drug_screen.columns))
+        #
+        # train_drug_screen.drop(common_columns[1:100], axis=1, inplace=True)
+        # test_drug_screen = test_drug_screen[common_columns[1:100]]
+
 
     # Step 4: Prepare input data for training
     x_cell_train, x_drug_train, y_train, cell_sizes, drug_sizes = RawDataLoader.prepare_input_data(train_data,
@@ -226,4 +231,4 @@ if __name__ == '__main__':
     torch.manual_seed(RANDOM_SEED)
     random.seed(RANDOM_SEED)
     np.random.seed(RANDOM_SEED)
-    run(30, is_test=False)
+    run(10, is_test=True)

utils.py

@@ -31,7 +31,7 @@ SIM_KERNEL = {'cell_CN': ('euclidean', 0.001), 'cell_exp': ('euclidean', 0.01),
 SAVE_MODEL = False  # Change it to True to save the trained model
 VARIATIONAL_AUTOENCODERS = False
 # DATA_MODALITIES=['cell_CN','cell_exp','cell_methy','cell_mut','drug_comp','drug_DT'] # Change this list to only consider specific data modalities
-DATA_MODALITIES = ['cell_exp', 'drug_desc','drug_finger']
+DATA_MODALITIES = ['cell_CN','cell_exp','cell_methy','cell_mut', 'drug_desc','drug_finger']
 RANDOM_SEED = 42  # Must be used wherever can be used