Usage¶

Here we describe the general usage of GNNSubNet.

Getting Started¶

To use GNNSubNet, the following general procedure is followed:

Create a GNNSubNet object which points to a directory containing the data you wish to analyse
Train a graph neural network using this data
Run the explainer to explain the results

In the following subsections we will go through each of these steps in detail.

Step 1: Create a new GNNSubNet object¶

First we import the main GNNSubNet module as gnn, and then create a new GNNSubNet object, which we will call g. The datasets can be found here: https://github.com/pievos101/GNN-SubNet/tree/main/TCGA

from GNNSubNet import GNNSubNet as gnn

# location of the files
loc   = "/home/bastian/GitHub/GNN-SubNet/TCGA"
# PPI network
ppi   = f'{loc}/KIDNEY_RANDOM_PPI.txt'
# single-omic features
#feats = [f'{loc}/KIDNEY_RANDOM_Methy_FEATURES.txt']
# multi-omic features
feats = [f'{loc}/KIDNEY_RANDOM_mRNA_FEATURES.txt', f'{loc}/KIDNEY_RANDOM_Methy_FEATURES.txt']
# outcome class
targ  = f'{loc}/KIDNEY_RANDOM_TARGET.txt'

# Load the multi-omics data
g = gnn.GNNSubNet(loc, ppi, feats, targ)
g.summary()

Once we have created our GNNSubNet object g we can train.

Step 2: Train the GNN classifier¶

# Train the GNN classifier and validate performance on a test set
g.train()

# Check the performance of the classifier
g.accuracy
g.confusion_matrix

Step 3: Run the exlainer on the trained network¶

# Run the Explainer with 4 iterations (10 is recommended)
g.explain(3)

# Edge and Node (Gene) Importances
g.gene_names
g.edges

# Importances
g.edge_mask
g.node_mask

# Detected modules and their importances
g.modules
g.modules[0]
g.gene_names[g.modules[0]]

g.module_importances