Ignoring stratigraphic age uncertainty leads to erroneous estimates of species divergence times under the fossilized birth–death process

Abstract

Fossil information is essential for estimating species divergence times, and can be integrated into Bayesian phylogenetic inference using the fossilized birth–death (FBD) process. An important aspect of palaeontological data is the uncertainty surrounding specimen ages, which can be handled in different ways during inference. The most common approach is to fix fossil ages to a point estimate within the known age interval. Alternatively, age uncertainty can be incorporated by using priors, and fossil ages are then directly sampled as part of the inference. This study presents a comparison of alternative approaches for handling fossil age uncertainty in analysis using the FBD process. Based on simulations, we find that fixing fossil ages to the midpoint or a random point drawn from within the stratigraphic age range leads to biases in divergence time estimates, while sampling fossil ages leads to estimates that are similar to inferences that employ the correct ages of fossils. Second, we show a comparison using an empirical dataset of extant and fossil cetaceans, which confirms that different methods of handling fossil age uncertainty lead to large differences in estimated node ages. Stratigraphic age uncertainty should thus not be ignored in divergence time estimation and instead should be incorporated explicitly.

Incorporating uncertainty is essential to macroecological inferences: Grass, grit, and the evolution of kangaroos

“…These divergence estimates differ considerably from analyses implementing fixed fossil ages (as minimum, mean, or maximum stratigraphic bounds), including another tip-dating study (Cascini et al 2018), and generally fall between estimates from mean and maximum fixed 380 dates (Fig.2). This exercise suggests that signal in the morphological and molecular clocks can contribute to fossil age information, and is consistent with other recent study in this area suggesting that fixing tip ages should be avoided (Barido-Sottani et al 2019). Though it is important to note that in divergence dating analyses focused on intraspecific sampling implementing a strict molecular clock, divergence estimates may not differ between fixed 385 and prior-informed tip ages (Molak et al 2013).…”

“…This provides the opportunity to co-estimate the phylogeny and divergence times, while providing structured 130 priors on nodes which may otherwise be driven to unrealistic deep or shallow values. One shortcoming of nearly all implementations of tip-dating however, is the requirement of fixing fossil ages to a single value (Heath et al 2014;Barido-Sottani et al 2019). Except where radiometrically dated, fossil age estimates are rarely precise enough to fit this expectation, and so we often arbitrarily use the median value or a bio-correlated guess within a fossil's 135 age interval.…”

“…Except where radiometrically dated, fossil age estimates are rarely precise enough to fit this expectation, and so we often arbitrarily use the median value or a bio-correlated guess within a fossil's 135 age interval. This practice can lead to biased inferences when values nearer maximum or minimum ages are consistently applied, or when an age is randomly assigned (Barido-Sottani et al 2019). Fixed fossil ages also ignore the persistence of lineages for potentially long periods of time, summarizing an extinct taxon to a single point estimate.…”

Phylogenomics of monitor lizards and the role of competition in dictating body size disparity

“…In most implementations of tip-dating fossil ages are 540 fixed to a single value-most often this is the median value between upper and lower bounds. To avoid unintentional bias in choosing exact fossil ages, we instead incorporate uncertainty by sampling from informed uniform priors allowing the fossil ages to be jointly estimated (Barido-Sottani et al 2019). Morphological data were modelled under the Mkv model, a special case of the Mk model (Lewis 2001)-the most commonly used model for discrete morphological data.…”

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

“…Barido-Sottani et al (2019a) focused on one particular aspect of the fossil record, namely the uncertainty associated with the age assigned to each fossil sample. As the age of fossils is established in reference to the geological record, fossil samples are not dated to a single value but rather to an interval of time; this is referred to hereafter as the “age range” of the sample.…”