PHYLOGENETIC PALEOBIOLOGY AS A TOOL FOR TESTING MACROEVOLUTIONARY HYPOTHESES: CASE STUDIES WITH CAMBRIAN ECHINODERMS (Invited Presentation)

It is often stated that a great advantage of paleontology among the evolutionary sciences is its temporal dimension. Indeed, time plays a crucial role in understanding virtually all macroevolutionary patterns, including phylogenies of fossil taxa. Macroevolutionary inference in paleobiology has historically been a two-step process involving phylogeny inference followed by downstream tests of alternative macroevolutionary models. However, the ability to estimate patterns of trait change, lineage diversification, and phylogenetic relationships are not generally independent of one another. To help overcome these issues, phylogenetic paleobiologists have increasingly been interested in statistical approaches that attempt to more fully account for the idiosyncrasies of fossil data.

Modern Bayesian tip-dating methods incorporating the Fossilized Birth-Death process involve simultaneous inference of tree topology, divergences, and other macroevolutionary parameters. Unfortunately, these models are highly complex, and decisions regarding model configurations are often left to a series of computationally expensive analyses to estimate marginal likelihoods and compute Bayes Factors, which are used to evaluate evidence for different evolutionary models.

Here, we apply a reversible jump Markov-chain Monte Carlo approach to investigate macroevolutionary dynamics in an enigmatic clade of Cambrian echinoderms, the Cincta, which have become a model group for phylogenetic paleobiology. In the spirit of simultaneous inference, we use rjMCMC to jointly infer a dated phylogeny of cinctan species while also directly testing alternative hypotheses of cinctan evolution. For example, we compete multiple candidate character models, including scenarios where rates differ between feeding traits, functional vs. non-functional characters, and even among different morphological modules of the cinctan body plan. Similarly, we test for temporal shifts in species diversification while simultaneously allowing for different clock models of character evolution describing rate variation throughout their phylogeny. Finally, we show how the posterior distribution of cinctan phylogenies themselves can be use to make claims about their evolutionary relationships, patterns of ancestor—descendant pairs, and the tempo and mode of speciation in the fossil record.