AN ABC APPROACH TO THE CO-ESTIMATION OF TAPHONOMIC PARAMETERS AND EVOLUTIONARY RATES

A suite of techniques for examining trait-dependent speciation and extinction rates (BiSSE, MuSSE, QuaSSE, and GeoSSE) have revolutionized how biologists interpret phylogenies of extant organisms. These techniques require the known proportion of taxa included in a phylogeny, require random sampling, and exclude information on extinct sister clades. Here I present an Approximate Bayesian Computation (ABC) approach for the construction and comparison of different models that include taphonomic and evolutionary parameters. The ABC framework makes use of summary statistics, numerous computer simulations, and distances between the observed and simulated summary statistics to approximate a posterior distribution for models. ABC does not require the calculation of likelihoods; this provides a very flexible framework for working with complex models for which the likelihood functions are unknown or not computable. This approach can include all phylogenetic information about extinct sister clades, works on clades that are entirely extinct, and can account for sampling that is correlated with a character of interest (e.g. size, biogeographic region, preferred environment). Several models for the Pleurocystitinae, an extinct clade of Paleozoic echinoderms, are compared.