Chapter meeting -
Colloquium talk
Colloquium talk
Click here to RSVP by March 24.
Free of charge.
Free parking and food will be provided.
Participation in our chapter activities is not restricted to ASA members.
The chapter meeting will be held at FO 1.202, UT Dallas. It will also be live-streamed through Microsoft Teams via this link.
The parking permit is available upon request. Click here to access the interactive UT Dallas campus map.
Factor models have been popular for efficiently inferring relationships among multivariate continuous responses using a small number of parameters. Motivated by complex count data from microbiome studies using next-generation sequencing, we develop a sparse Bayesian factor model for count table data to capture interactions among microorganisms. Our model employes a rounded kernel mixture model using a Dirichlet process (DP) prior with log-normal mixture kernels for count vectors. A factor model is used to model the covariance matrix of the mixing kernel that describes microorganism interaction. We construct a Dirichlet-Horseshoe (Dir-HS) shrinkage prior and use it as a joint prior for factor loading vectors. Joint sparsity induced by a Dir-HS prior greatly improves the performance in high-dimensional applications. Building on this sparse factor model, we introduce two extensions for more complex settings: (1) a sparse Bayesian group factor model (Sp-BGFM), developed to infer relationships among microbes in different domains, and (2) a Bayesian covariate-dependent sparse factor model, developed to infer covariate-dependent interdependencies among microbes. We thoroughly investigate the theoretical properties of our models and conduct extensive simulation studies to evaluate their performance. Real data examples from microbiome studies are used for illustration.
Joint work with Shuangjie Zhang (Statistics, Texas A&M), Yuning Shen (Chemical and Biomolecular Engineering, UCLA), Irene A. Chen (Chemical and Biomolecular Engineering, UCLA), Michael Patnode (Microbiology and Environmental Toxicology, UCSC)
Dr. Juhee Lee is a professor of Statistics at the University of California Santa Cruz. She received her Ph.D. in Statistics at The Ohio State University in 2010. Her research interests include Bayesian statistical modeling, Bayesian nonparametrics, biological applications, clinical trial designs, statistical decision making, and survival analysis.
Organizing committee
Qiwei Li (chair)
Hyunwoong Chang (co-chair)
Judging committee
Xiwei Tang (chair)
Kevin Lutz (co-chair)
Supporting staffs
James Smith
Gabriella Taylor