Title: Bayesian multi-layered Gaussian graphical models
Authors: Min Jin Ha - UT MD Anderson Cancer Center (United States) [presenting]
Francesco Stingo - University of Florence (Italy)
Veerabhadran Baladandayuthapani - UT MD Anderson Cancer Center (United States)
Abstract: Simultaneous modeling of data arising from multiple ordered layers provides insight into a holistic picture of the interactive system and the flow of information. Chain graphs have been used to model the layered architecture of networks where the vertices can be naturally partitioned into ordered layers that exhibit undirected and directed acyclic relations within and between the layers. We use the multi-layered Gaussian graphical model (mlGGM) to describe a conditional independence structure on a chain graph and propose a Bayesian node-wise selection (BANS) method that coherently accounts for dependencies in the mlGGM. Using variable selection priors for each of the node-wise regressions allows for flexible modeling and the incorporation of edge-specific prior knowledge. Through simulation data generated from various mlGGMs, we demonstrate that our node-wise regression method outperforms other related multivariate regression-based methodologies. We apply BANS to identify integrative networks for key signaling pathways in kidney cancer and dynamic signaling networks using longitudinal protein data from a breast cancer cell line.