Example 3: Disease Prediction Using Graph Information (SmCCNet + DPMON)
=======================================================================

This example demonstrates a hybrid workflow combining SmCCNet for network generation and DPMON for disease prediction using multi-omics data and graph information.

**Workflow Overview:**

1. **Network Construction (SmCCNet):**
   Generates an adjacency matrix from multi-omics data using `SmCCNet`. The resulting matrix represents relationships between features.

2. **Disease Prediction (DPMON):**
   Utilizes the generated adjacency matrix to predict disease phenotypes using the `DPMON` model.

**Step-by-Step Guide:**

1. **Setup Input Data:**
   - Prepare the following inputs:
   
     - **Proteins Data (`omics_proteins`)**: Pandas DataFrame containing protein features.
     - **Metabolites Data (`omics_metabolites`)**: Pandas DataFrame containing metabolite features.
     - **Phenotype Data (`phenotype_data`)**: Pandas Series with phenotype labels.
     - **Clinical Data (`clinical_data`)**: Pandas DataFrame with clinical features.

   Example input data structure:

   .. code-block:: python

      omics_proteins = pd.DataFrame({
          'protein_feature1': [0.1, 0.2],
          'protein_feature2': [0.3, 0.4]
      }, index=['Sample1', 'Sample2'])

      omics_metabolites = pd.DataFrame({
          'metabolite_feature1': [0.5, 0.6],
          'metabolite_feature2': [0.7, 0.8]
      }, index=['Sample1', 'Sample2'])

      phenotype_data = pd.Series([1, 0], index=['Sample1', 'Sample2'])

      clinical_data = pd.DataFrame({
          'clinical_feature1': [5, 3],
          'clinical_feature2': [7, 2]
      }, index=['Sample1', 'Sample2'])

2. **Run SmCCNet to Generate the Adjacency Matrix:**

   Use `SmCCNet` to construct the network from multi-omics data:

   .. code-block:: python

      smccnet_instance = SmCCNet(
          phenotype_data=phenotype_data,
          omics_data=pd.concat([omics_proteins, omics_metabolites], axis=1),
          data_types=['protein', 'metabolite'],
          kfold=5,
          summarization='PCA',
          seed=732,
      )
      adjacency_matrix = smccnet_instance.run()
      print("Adjacency matrix generated using SmCCNet.")

   This step produces the adjacency matrix, which encodes the relationships between features in your data.

3. **Run DPMON for Disease Prediction:**

   With the adjacency matrix from SmCCNet, use `DPMON` to predict disease phenotypes:

   .. code-block:: python

      dpmon_instance = DPMON(
          adjacency_matrix=adjacency_matrix,
          omics_list=[omics_proteins, omics_metabolites],
          phenotype_data=phenotype_data,
          features_data=clinical_data,
          model='GCN',  
          tune=False,  
          gpu=False     
      )

      predictions_df = dpmon_instance.run()

      if not predictions_df.empty:
          print("DPMON workflow completed successfully. Predictions generated.")
      else:
          print("DPMON hyperparameter tuning completed. No predictions were generated.")

4. **Save and Interpret Results:**

   The predictions generated by `DPMON` are returned as a DataFrame, providing insights into disease associations based on the integrated omics data.

   .. code-block:: python

      print("DPMON Predictions:")
      print(predictions_df)

**Running the Example:**

To execute the complete workflow:

.. literalinclude:: ../examples/example_3.py
   :language: python
   :caption: Complete SmCCNet + DPMON Hybrid Workflow.

**Result Interpretation:**

- **Adjacency Matrix**: Represents the constructed network from multi-omics data, indicating the strength and presence of relationships between features.
- **Disease Predictions**: Provides predictions related to disease states based on the integrated omics and network data, facilitating biomarker discovery and patient stratification.