Joint Estimation of Gaussian Graphical Models in Spatial and Temporal Data

• Activity Number: 68
• Type: Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
• Sponsor: ENAR
• Author(s): Zhixiang Lin*+
• Keywords: Gaussian Graphical Model ; Temporal and spatial data ; Neurodevelopment ; Network estimation

Abstract

Human neurodevelopment is a highly regulated biological process. In this presentation, we study the dynamic changes of neurodevelopment through the analysis of human brain microarray data sampled from 16 brain regions in 15 time periods of neurodevelopment. The biological networks are captured by Gaussian Graphical Models (GGMs). We propose to jointly estimate the GGMs in each time period and in each brain region. A penalized likelihood model is used to efficiently utilize the information embedded in the brain region similarity and temporal dependency in our approach. We also develop and implement an efficient algorithm to estimate the precision matrices. Simulation studies suggest that our approach achieves lower misclassification error and potential gain in power compared with models not incorporating spatial similarity and temporal dependency.