Phylogenetic tree space exploration in online Bayesian phylodynamic inference

Available approaches to perform online Bayesian phylodynamic inference support a rather limited set of models and need to be further developed to enable a wide range of scenarios under different (coalescent and molecular clock) models. In a first part of my PhD, I will assess the exploration of phylogenetic tree space in standard Bayesian phylodynamic inference via novel methods known as tree effective sample size (ESS) measures, starting with a simulation study in which I can control the Monte Carlo error and devise a useful set of guidelines for researchers performing these types of analyses. Based on my findings, I will evaluate the performance of these tree ESS metrics on highly relevant data sets of HIV, SARS-CoV-2 and Ebola virus that have been published in recent years. Further, the impact of phylogenetic placement approaches on the thoroughness of the phylogenetic tree search space exploration in online Bayesian phylodynamic inference is currently unknown and needs to be carefully assessed. The main goal of this work is to carefully assess how thorough Bayesian phylodynamic analyses have been over the past years in correctly capturing phylogenetic uncertainty and their overall Monte Carlo error, to advise researchers on how and how long to run their Bayesian phylodynamic analyses.