High-Dimensional Bayesian Model Averaging for Gene Network Inference

• Activity Number: 400 - Multiple Aspects of Bayesian Model Selection and Variable Selection in Linear and Nonlinear Models
• Type: Topic Contributed
• Date/Time: Wednesday, August 5, 2020 : 1:00 PM to 2:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Adrian Raftery*
• Companies: University of Washington
• Keywords: human genome data; knockdown data; prior probabilities; supervised learning; gene ontology; genetic regulatory network

Abstract

Inferring gene networks from large-scale human genome data is hard in part due to the high dimension of the search space. We present a computationally efficient Bayesian model averaging approach, integrating external data with knockdown data from human cell lines. We use multiple data sources, including gene expression, genome-wide binding, gene ontology, and known pathways, and we use a supervised learning method to compute prior probabilities of regulatory relationships. We show that our integrated method improves the accuracy of inferred gene networks. Joint work with Xiao Liang, William C. Young, Ling-Hong Hung and Ka Yee Yeung.