Multi-Omics Enhanced Pathway Graph Neural Networks for Personalized Cancer Prognosis

By representing biological pathways as graphs and incorporating multi-omics features as nodes for these graphs, this thesis investigates the application of graph neural networks (GNNs) for cancer prognosis. By integrating mutation data into pathway-aware models obtained from gene expressions, the goal is to enhance patient risk stratification. To enable more individualized and interpretable predictions, the suggested framework expands upon and improves upon current models like PathGNN.