This document summarizes a study that integrates gene expression and DNA methylation data to predict survival in breast cancer patients. The study constructs an integrated gene-gene graph incorporating pathway information and gene interactions. It then uses a directed random walk approach to infer pathway activities from the graph. A denoising autoencoder is applied to select important pathways related to breast cancer survival. The top pathways identified include dorso-ventral axis formation and neurotrophin signaling. Analysis found differentially expressed genes and methylation features in the pathways are associated with breast cancer. The integrated approach improved cancer classification performance over using single data types.