Date/Time
Date(s) - 12/02/2026
1:30 pm - 2:30 pm

Speaker: Ayesha Ali (University of Guelph)

Location: Hamilton Hall, Room 217

Title: Dominating Hyperplane Regularization for Multivariate Count Data: Application to the Gut Microbiome

Abstract: Multivariate count regression involves simultaneously associating multiple covariates with the observed count vector across several response categories. For gut microbiome data from US immigrants, interest may lie in how acculturation is associated with gut bacterial abundances. Bacterial counts or sequenced OTU counts are typically associated through a taxonomic tree so the standard Dirichlet-multinomial (DM) distribution, which conditions on a particular taxonomic level, may be too restrictive for certain applications.  However, the Dirichlet-Tree multinomial distribution relaxes this restriction but, like the DM distribution, falls outside the exponential family.  When performing variable selection, complex penalty functions, such as the sparse group lasso where coefficients for each covariate across the response categories are grouped together, optimization of the objective function is further challenged.  Here, we introduce Dominating Hyperplane Regularization (DHR) as a stable method of optimization. Based on the majorization-minimization framework, a suitable surrogate for the penalty function is used during optimization. We also introduce novel taxonomic tree-guided penalty functions that leverage known relationships among taxa. We study the performance of DHR for these novel models based on real world data from the Hispanic Community Health Study/Study of Latinos.