Topic: RevBayes: A program for Bayesian inference of phylogeny using probabilistic graphical models and an interpreted language
Dates: August 7-11, 2017
This course featured RevBayes, an exciting new program for Bayesian inference of phylogeny. RevBayes is the successor to the popular program MrBayes, but represents both a complete rewrite of the computer code and a fundamental re-conception of phylogenetic models. Specifically, RevBayes adopts a 'graphical-model' framework that views all statistical models as comprised of components that can be assembled in myriad configurations to explore a corresponding array of questions. This graphical-model approach to phylogenetic inference provides effectively infinite flexibility. Moreover, the graphical models are specified using an R-like language, Rev, that empowers users to construct arbitrarily complex phylogenetic models from simple component parts (i.e. random variables, parameter transformations and constants of different sorts).
This course was focused on phylogenetic trees and comparative-phylogenetic methods, including divergence-time estimation, morphological evolution, lineage diversification, and historical biogeography. Instruction was based on a combination of carefully tailored lectures introducing the theoretical and conceptual basis of each inference problem and hands-on computer tutorials demonstrating how to explore these questions using RevBayes (see http://revbayes.github.io/tutorials.html).
Participants were not assumed to have expertise in phylogenetic theory; rather, we provided an accessible introduction to Bayesian statistical inference and stochastic models. We assumed only that participants were familiar with phylogenetic trees and their applications to evolutionary biology. We therefore anticipated that this course would be most suitable for senior PhD students, postdoctoral researchers, and faculty who wanted to learn these techniques.
If you are interested in partnering with NIMBioS to drive a new workshop or tutorial, please visit our partner page for more information.