Network Propagation for the Analysis of Multi-omics Data

Network propagation has been used as the state-of-the-art network analysis method which can overcome the limitations of the existing methods. The area where network propagation truly shines is the application to multi-omics data. It integrates interactome data with other omics data to be utilized for further analyses. In multi-omics data, there are multiple levels of complexity for which network propagation techniques can be used effectively. Here, we categorized recent network propagation research works in bioinformatics into three levels: DNA, RNA, and pathway levels. At the DNA level, network propagation is applicable for genome data analysis. In particular, for the analysis of genome-wide mutation profiles in cancer, network propagation has been successfully used to detect subnetworks, driver mutation profiles, and cancer subtypes. At the RNA-level analysis of transcriptome data, gene expression information can be propagated to obtain valuable condition-specific information such as the ranking of how genes react to certain environments. Such derived information can be utilized in differentially expressed gene detection, gene prioritization, and hypothesis testing. Transcriptome data can be analyzed at the biological pathway level by utilizing network propagation techniques. With graph convolutional network techniques, network propagation can be exploited for a disease subtype classification problem where pathway information is used to characterize biological mechanisms of a disease.