Biostatistics Seminar: BST 290: Sebastian Hohna, UC Davis
DATE: Tuesday, November 4, 2014
LOCATION: 1147 Math Sciences Building
SPEAKER: SEBASTIAN HOHNA, Post-doctoral Fellow, Dept of Evolution and Ecology, UC Davis
Title: “Estimating mass-extinction events from molecular phylogenies and joint Bayesian inference of phylogeny, diversification rates and divergence times”
Abstract: Species diversity is modeled by a birth-death process with lineage-homogeneous but time-heterogeneous speciation and extinction rates and explicitly modeled mass-extinction events. This enables diversification rates and past diversifications event —such as mass-extinction events— to be estimated (in principle) from molecular phylogenies. However, incomplete taxon sampling, especially non-uniform taxon sampling, biases our inference results.
Here, I will give a brief overview of the birth-death process with lineage-homogenous and time-heterogeneous speciation and extinction rates and how this process can be extended to include different scenarios of incomplete taxon sampling. Furthermore, I will present a hierarchical Bayesian inference using reversible jump MCMC that is able to infer past mass-extinction events from molecular data only.
The shortcomings of the current method is that is relies on a perfectly known phylogeny without uncertainty. I will present in an outlook how our new software package RevBayes, that implements a probabilisitic graphical model approach, is be able to deal with such analysis using a joint approach for estimating phylogeny, substitution rate parameters, diversification rate, etc.