Phylogenetic Inference from an Incomplete Fossil Record

Poster for Congress of the European Society of Evolutionary Biology (ESEB) 2025 in Barcelona, Spain.

Abstract text:

Morphological and temporal data from fossils has become an increasingly important part of phylogenetic inference. It is a crucial part of the construction of time trees, which open the opportunity to test a wide range of evolutionary hypotheses via comparative methods.

However, the fossil record is known to be incomplete: Organisms might not fossilize, and fossils might be destroyed by erosive and taphonomic processes. However, gaps in the fossil record are not random, but can be predictably linked to external drivers of stratigraphic architectures such as changes in sea level.

We explore effect of the incompleteness of the fossil record on the precision and accuracy of phylogenetic inference. For this, we combine forward models of the rock record with simulations of the fossilized birth-death model to generate a synthetic fossil record. We explore how well the true evolutionary history (tree topology and model parameters) can be recovered from stratigraphic architectures controlled by varying external drivers. Our results will elucidate how data sparsity introduced by geological effects influence tree inference and identify scenarios that are best suited for the recovery of evolutionary change from fossil data.