PhD: Reproducibility of morphological phylogenetics

Evolutionary trees are an essential tool for testing evolutionary hypotheses and building the history of life on earth. Whilst the science of building trees (phylogenetics) is mature, the principles of “open data” have rarely been applied consistently. The reproducibility of phylogenies generated from genomic data has been investigated, but the same is not true of trees generated from morphological data. This represents a challenge to morphological phylogenetics as a discipline, especially considering that there is anecdotal evidence that the results of numerous studies are not reproducible. How confident can we be in the evolutionary inferences we make from anatomical and fossil data? What is the scale of these problems, and are they evenly distributed across different parts of the tree of life, journals, data types or even authors? We propose to directly test the reproducibility of morphological phylogenies for the first time by taking a “big-data” meta-analytical approach. Alongside testing for reproducibility we will also be testing the effect of applying different methods of phylogenetic inference to multiple datasets. We will assess how different empirical topological results are in the context of debate about the accuracy of parsimony versus Bayesian inference. Ultimately, we aim to test the “fitness” of the field of morphological phylogenetics and put in place recommendations and guidelines for the future of the discipline.