Practical guide and review of fossil tip-dating in phylogenetics

Phylogenetic tip-dating has been and still is revolutionizing evolutionary biology in several ways. Tip-dating, where fossils are placed into a phylogeny as tips based on molecular and/or morphological character information, provides a more principled approach to infer time-calibrated phylogenies compared with node-dating. Additionally, phylogenetic trees with fossils as tips become more and more important to elucidate evolutionary processes in macroevolutionary studies, e.g., deciphering diversification patterns and directional phenotypic evolution. Tip-dating is slowly gathering popularity in empirical applications and has progressed substantially since its first demonstration in 2011, with respect to improved statistical models, software, and datasets. Nevertheless, executing a phylogenetic tip-dating analysis is complicated and comes with many challenges. Here, we provide an extensive review and overview of methods and models for phylogenetic tip-dating analyses. We focus both on data collection and preparation as well as on modeling choices. We start with a survey of all published phylogenetic tip-dating studies to date, showing common data and modeling choices as well as trends towards new approaches. Then, we walk readers through sections of molecular evolution, morphological evolution (both for discrete and continuous data), and lineage evolution (the fossilized-birth-death process). In each section, we describe the data and standard models with their underlying assumptions, and provide an outlook and practical recommendations.

Methods
We conducted a literature survey in order to gather information about tip-dating analyses on empirical datasets. We used Google Scholar to look at all publications citing a selection of foundational tip-dating studies, and among those, we identified publications performing tip-dating analyses on empirical datasets. We compiled relevant information---including software of choice, dataset size, types of data used, and clock and tree models employed. When it was possible, this information was extracted directly from data and script files provided as supplementary material of each publication; otherwise, we used the information written in the methodological sections of the main and supplementary texts.

While we originally included tip-dating studies without morphological data in which either the whole topology was fixed or topological constraints were applied to extinct taxa, we pruned those studies from our survey before downstream analyses for summary figures and statistics. This is because we wanted to focus on studies that use tip-dating as an analytical approach to co-estimate the topological position of extinct taxa with divergence times, rather than just estimating divergence times on an a priori fixed topology, or pruning extinct tips afterwards to obtain a dated phylogenetic tree of only extant species.