DIVERSIFICATION TRENDS INFERRED FROM THE FOSSIL RECORD

Macroevolution focuses on patterns and processes occurring above the level of species and over geological timescales (Raia, 2016). Investigating diversification processes, both morphological and taxonomical, gives a chance to answer important questions in evolutionary biology. Why do some clades have more species than others? Why do some groups undergo striking adaptive radiations, but others persist for millions of years as living fossils? Why do some groups have much more ecological or morphological diversity than others? Does anything limit the number of species on earth and what? These complex questions share a common underlying feature: all, to some degree, concern rates of macroevolutionary change that occur across geological timescales (Rabosky & Slater, 2014). The aim of my project was to produce a coherent array of new methods to investigate phenotypic and taxonomic diversification by using phylogenies including extinct species. I started by developing RRphylo, a new phylogenetic comparative method based on phylogenetic ridge regression, which works with a phylogenetic tree and phenotypic data (univariate or multivariate either) to estimate branch-wise rates of phenotypic evolution and ancestral characters simultaneously. The main innovations, which translate in advantages of RRphylo over existing methods, lies in the absence of any a priori hypothesis about the mode of phenotypic evolution and in its ability to deal with fossil phylogenies. Both these factors make RRphylo very suitable to study phenotypic evolution in its different facets. I further implemented RRphylo to locate clade- or state-related shifts in absolute rates of phenotypic evolution, to integrate the effect of additional (to the phylogeny) predictors on rates estimation, to identify temporal trends in phenotypic mean and evolutionary rates occurring on the entire tree or pertaining individual clades, to identify instances of morphological convergence, to include ancestral character information derived from the fossil record, and to work with discrete variables. All these tools are collected into the RRphylo R package, online from April 2018, and counting > 14000 downloads on CRAN to date. I have been handling the maintenance and updates of the RRphylo package for both the release (https://cran.r-project.org/web/packages/RRphylo/index.html) and development (https://github.com/pasraia/RRphylo) versions, and creating/updating explanatory vignettes to facilitate its usage