Abstract:
|
In cancer research, multidimensional studies are gaining popularity. In such studies, multiple types of genetic, genomic, and epigenetic measurements are collected on the same subjects. Many of the existing studies are limited by ignoring the interconnections among (epi)genetic units. (Epi)Genetic units of the same type and different types have complex regulation relationships. In this study, we propose describing such relationships using a weighted network structure, which can be constructed using regularized estimation. Based on the network, we can prioritize nodes (genes, methylation loci, etc.), identify modules and hubs, as well as construct predictive models. The proposed methods are applied to the analysis of TCGA (The Cancer Genome Atlas) data.
|