Example 1: Identification of context-specific EGs

This example shows how to use HELP for computing two-class and three-class labelling of EGs based on tissue or disease-related information. The workflow involves the following steps

• 1. Install HELP from GitHub
• 2. Download the input files
• 3. Load the input files
• Example 1.1 two-class labeling of EGs based on tissue information
• Example 1.2 three-class labeling of EGs based on tissue information
• Example 1.3 two-class labeling of EGs based on disease-related information

Steps 1.-3. are needed only once, while steps 4.-5. are differentiated and executed three times to compute:

Example 2: Identification of uncommon context-specific EGs

This example shows how to use HELP for identifying uncommon tissue-specific EGs. The workflow involves the following steps

• 1. Install HELP from GitHub
• 2. Download the input files
• 3. Load the input file
• 4. Filter the information to be exploited
• 5. Compute EGs common to all tissues (pan-tissue labeling)
• 6. Subtract pan-tissue EGs from those of the chosen tissue

Finally, the example shows various ways of visualizing the obtained results.

Example 3: Prediction of EGs for a tissue

This example shows how to use HELP to estimate the performance of EG prediction for a tissue. The workflow involves the following steps

• 1. Install HELP from GitHub
• 2. Download the input files
• 3. Load the input files and process the tissue attributes
• 4. Estimate the performance of EGs prediction
• 5. Compute TPR for ucsEGs and csEGs

Step 5. shows how to compute the True Positive Rate (TPR) for ucsEGs and csEGs and show their bar plot.

Example 4: Extraction of network embedding from tissue PPI

This example shows how to use in the HELP framework the graph embedding functions of the Karateclub <https://karateclub.readthedocs.io/> python package to accomplish node embedding on the Protein-Protein Interaction (PPI) network of a tissue. The results can be used as features for EG prediction with HELP. The workflow involves the following steps

• 1. Install HELP from GitHub (and Karateclub)
• 2. Download the PPI input file
• 3. Load the PPI input file
• 4. Compute the embedding
• 5. Save and show the embedding

Example 5: Reproduce the experiments reported in the HELP paper

This example shows how to use HELP to reproduce the experiments reported in the HELP paper. The workflow involves the following steps

• 1. Install HELP from GitHub
• 2. Download the input files
• 3. Download the script for the experiments and show the man page
• 4. Run the E vs NE experiments
• 5. Run the E vs sNE experiments