Update 3rd stage

3-TrainMogonet/README.md

@@ -12,4 +12,8 @@ git clone https://github.com/txWang/MOGONET.git
 
 3. Run `main_mogonet.py` to train the network and store the model
 
-4. Run `main_biomarker.py` to evaluate the feature importances and extract biomarkers from the model.
+4. Run `main_biomarker.py` to evaluate the feature importances and extract biomarkers from the model.
+
+### Alternative approach
+
+Or you can configurate the `runner.py` code and run it.

3-TrainMogonet/runner.py

@@ -0,0 +1,31 @@
+import os
+import copy
+
+from mogonet.feat_importance import cal_feat_imp, summarize_imp_feat
+from mogonet.train_test import train_test
+
+data_folder = 'ICGC'
+view_list = [1, 2]
+
+
+def run_mogonet(num_class):
+    os.chdir('./mogonet')
+    num_epoch_pretrain = 500
+    num_epoch = 2500
+    lr_e_pretrain = 1e-3
+    lr_e = 5e-4
+    lr_c = 1e-3
+    train_test(data_folder, view_list, num_class,
+               lr_e_pretrain, lr_e, lr_c,
+               num_epoch_pretrain, num_epoch)
+
+
+def run_biomarker(num_class):
+    os.chdir('./mogonet')
+    model_folder = os.path.join(data_folder, 'models')
+    featimp_list_list = []
+    for rep in range(5):
+        featimp_list = cal_feat_imp(data_folder, os.path.join(model_folder, str(rep + 1)),
+                                    view_list, num_class)
+        featimp_list_list.append(copy.deepcopy(featimp_list))
+    summarize_imp_feat(featimp_list_list)