Assistant Professor of the Practice of Statistical Science
How quickly is an infectious pathogen’s epitope mutating? How fast is said pathogen diffusing across a geographic landscape? What factors contribute to, or otherwise augment, transmissibility?
My current research focuses on leveraging Bayesian phylogenetics to answer questions like these that lie at the interface of epidemiology, genetics and evolution. Within this framework, one is often interested in estimating thousands of highly correlated parameters that describe complex, hierarchical data generative processes. Typically, the number of parameters grows with the data. To study increasingly massive data sets, including genomic sequence and spatial coordinate data, I build scalable statistical models together with scalable inference machinery and implement my work in the popular open source Bayesian Evolutionary Analysis Sampling Trees (BEAST) software package. Current methodologies of interest include Markov chain Monte Carlo sampling, Bayesian variable selection via shrinkage priors and dynamic programming algorithms.