Phylogenetic trees are an important tool for biologists to understand how biological entities evolve through time, and they are informed by data from molecular sequences of DNA or RNA. Tree space is complicated and high-dimensional, which makes it extremely hard to traverse. Moreover, it admits many representations, each with its strengths and weaknesses. These characteristics complicate the use of Markov Chain Monte Carlo (MCMC) methods for computing expectations with respect to probability measures defined on phylogenetic space. One possible strategy for performing and studying MCMC in phylogenetic space is to find lower-dimensional projections that preserve the Markov property, either exactly or approximately. In this talk, we will discuss how to perform such a dimension reduction in the space of rooted trees using lumpability or quasi-lumpability with respect to subtrees, also known as clades. We study Metropolis-Hastings and lazy random walks on the rooted subtree prune-and-regraft (rSPR) graph and give bounds on the lumping error and total variation with respect to clades.

Texto informado pelo autor.

* Os participantes dos seminários não poderão acessar às dependências da FGV usando bermuda, chinelos, blusa modelo top ou cropped, minissaia ou camiseta regata. O uso da máscara é facultativo, porém é obrigatória a apresentação do comprovante de vacinação (físico ou digital).