Title: Simultaneous clustering and estimation of networks in multiple graphical models
Authors: Gen Li - University of Michigan Ann Arbor (United States) [presenting]
Abstract: The standard Gaussian graphical models have been widely used to investigate the dependency structure among variables in a single population. As it is increasingly common to collect data from multiple heterogeneous populations, multi-layered networks have become prevalent. We consider the simultaneous clustering and estimation of multiple graphical models. We build upon the Gaussian graphical models and utilize a sparse tensor decomposition approach to simultaneously cluster populations and estimate the underlying network structures among variables in each population. A penalized likelihood method is used to devise an alternating direction method of multipliers algorithm to estimate model parameters. We demonstrate the efficacy of the proposed method with comprehensive simulation studies. The application to the GTEx multi-tissue gene expression data provides important insights into tissue clustering and gene co-expression patterns in different tissues.