Ignoring fossil age uncertainty leads to inaccurate topology and divergence times in time calibrated tree inference

Abstract

15Time calibrated trees are challenging to estimate for many extinct groups of species due to the 16 incompleteness of the rock and fossil records. Additionally, the precise age of a sample is typically 17 not known as it may have occurred at any time during the time interval spanned by the rock layer. 18 1 Bayesian phylogenetic approaches provide a coherent framework for incorporating multiple sources 19 of evidence and uncertainty. In this study, we simulate datasets with characteristics typical of 20 Palaeozoic marine invertebrates, in terms of character and taxon sampling. We use these datasets 21 to examine the impact of different age handling methods on estimated topologies and divergence 22 times obtained using the fossilized birth-death process. Our results reiterate the importance of 23 modeling fossil age uncertainty, although we find that the overall impact of fossil age uncertainty 24 depends on both fossil taxon sampling and character sampling. When character sampling is low, 25 different approaches to handling fossil age uncertainty make little to no difference in the accuracy 26 and precision of the results. However, when character sampling is high, sampling the fossil ages 27 as part of the inference gives topology and divergence times estimates that are as good as those 28 obtained by fixing ages to the truth, whereas fixing fossil ages to incorrect values results in higher 29 error and lower coverage. Modeling fossil age uncertainty is thus critical, as fixing incorrect fossil 30 ages will negate the benefits of improved fossil and character sampling. 31 Introduction 32Estimating phylogenetic relationships and divergence times among species are key components of 33 piecing together evolutionary and geological history. Approaches to building time trees in paleo-34 biology have traditionally involved estimating the topology and branch lengths scaled to time in 35 separate, sequential analyses (Bapst and Hopkins, 2017). Bayesian phylogenetic models make it 36 possible to estimate these parameters in combination. An advantage of this joint inference is that 37 temporal evidence can be used to inform the tree topology, in combination with character data, 38 and the posterior output will better reflect the uncertainty associated with the results (Ronquist 39 et al., 2012).40Statistically coherent models for incorporating extinct species into time calibrated tree inference 41 only recently became available. In particular, the fossilized birth-death (FBD) process provides a 42 joint description of the diversification and fossil sampling processes (Stadler, 2010; Heath et al., 43 2014). Under this model, extinct dated samples are considered as part of the tree, therefore con-44 2 tributing temporal information, and their phylogenetic position can be recovered, either as terminal 45 branches (tips) or ancestral to other samples (sampled ancestors). This modeling framework has 46 created enormous potential for incorporating more paleontological data into divergence time analy-47 ses and we are on...

Convergent adaptation of true crabs (Decapoda: Brachyura) to a gradient of terrestrial environments

“…We randomly subsampled 10% of the 3,276 remaining occurrences, resulting in 328 occurrences. All fossil occurrences were assigned age ranges from the PBDB, each with a uniform distribution (Barido-Sottani et al 2019, 2020). To incorporate these fossil samples as part of the inferred evolutionary process, we used the unresolved time-homogeneous fossilized birth-death (FBD) tree prior (Stadler 2010; Heath et al 2014; O’Reilly and Donoghue 2020), conditioned on the same root prior as above.…”

Phylogenetic Comparative Methods: A User's Guide for Paleontologists

“…Perhaps you have previously estimated a phylogeny but now want to use it to answer new questions, perhaps you are interested in combining several smaller phylogenies to generate a supertree, or perhaps your specific question is not yet answerable in a Bayesian process-based framework (e.g., Warnock and Wright, 2020;Wright et al, 2020). The good news is that trees constructed in many different ways can be used in PCMs to make reliable inferences about trait diversification, provided that the tree is appropriately scaled to time using the stratigraphic record as an extra source of information (Bapst 2014a;Soul and Friedman 2015;Barido-Sottani et al 2020b).…”

Phylogenetic Comparative Methods: A User’s Guide for Paleontologists

“…Perhaps you have previously estimated a phylogeny but now want to use it to answer new questions, perhaps you are interested in combining several smaller phylogenies to generate a supertree, or perhaps your specific question is not yet answerable in a Bayesian process-based framework (e.g., Warnock and Wright, 2020;Wright et al, 2020). The good news is that trees constructed in many different ways can be used in PCMs to make reliable inferences about trait diversification, provided that the tree is appropriately scaled to time using the stratigraphic record as an extra source of information (Bapst 2014a;Soul and Friedman 2015;Barido-Sottani et al 2020b).…”