Incorporating paleontological data into phylogenetic inference can greatly enrich our understanding of evolutionary relationships by providing insights into the diversity and morphological evolution of a clade over geological timescales. Phylogenetic analysis of fossil data has been significantly aided by the introduction of the fossilized birth–death (FBD) process, a model that accounts for fossil sampling through time. A decade on from the first implementation of the FBD model, we explore its use in more than 170 empirical studies, summarizing insights gained through its application. We identify a number of challenges in applying the model in practice: it requires a working knowledge of paleontological data and their complex properties, Bayesian phylogenetics, and the mechanics of evolutionary models. To address some of these difficulties, we provide an introduction to the Bayesian phylogenetic framework, discuss important aspects of paleontological data, and finally describe the assumptions of the models used in paleobiology. We also present a number of exemplar empirical studies that have used the FBD model in different ways. Through this review, we aim to provide clarity on how paleontological data can best be used in phylogenetic inference. We hope to encourage communication between model developers and empirical researchers, with the ultimate goal of developing models that better reflect the data we have and the processes that generated them.
Reconstructing evolutionary relationships among organisms provides important insight into the history of life on Earth. Evolutionary (phylogenetic) trees can be used to show the relationships between extinct and extant organisms. By incorporating fossils, we can then estimate the timing of significant events such as a speciation event. The models used to generate trees in paleobiology combine different data sources, including molecular sequences from living organisms, the ages of fossils, and morphological information from both. Using this framework, we can learn about the rate of evolutionary dynamics of organisms, for example, the rate of speciation or diversification, or geographic movements through time. In this review article, we describe details of the statistical modeling framework used to integrate observations from living and fossil taxa. In particular, we focus on the use of the fossilized birth–death process, which is the only available model that allows us to include knowledge about the structure of the paleontological record into phylogenetic analyses. Because not all organisms and environments are equally well preserved, a flexible framework for working with fossils is essential for obtaining reliably dated phylogenies. In the decade since the model first became available in phylogenetic software, it has been used in more than 170 studies. We celebrate different applications of the model and highlight practical challenges. An important point that emerges from our discussion is that both the complexity of fossil data and details of the assumptions made by different models are crucial to consider. We hope to stimulate the exchange of ideas between researchers collecting and curating paleontological data and those developing models and software, with a view to further improving approaches to studying evolution in deep time.
