Overcoming the morphology problem of phylogenetics

Phylogenetic trees underpin reconstructions of evolutionary history and tests of evolutionary hypotheses. As such, they address major questions, such as, how and when did our modern biota come into being, and what is the shape and distribution of biodiversity and extinction? Advances in the acquisition and analysis of genetic sequence data have led to an increasing emphasis and reliance on molecular phylogenies, yet phenotypic evidence (morphology) remains vital. It enables us to make the link between organisms and their environment and thus demonstrate the mechanisms of evolutionary change. It is also the only way to include extinct taxa, and therefore provide a deep time perspective; fossils break up large gaps between the depauperate modern fauna, unlock sequences of evolutionary change (so called 'missing links') and provide a timescale for estimating evolutionary rates, including calibration of molecular clocks. Although morphology is acknowledged to be essential for phylogeny, it is also widely recognized as intrinsically problematic. Developmental and functional linkage can result in suites or modules of non-independent morphological characters and thus misleading patterns with respect to phylogeny reconstruction. These problematic phenomena are acknowledged to exist, but they have been largely ignored; morphological data are routinely taken at 'face value' and are treated as equivalent by both molecular and palaeontological studies. Furthermore, we have no idea about the distribution or influence of these phenomena. As such, our understanding of a range of evolutionary events is undermined, and our ability to reconstruct evolutionary history is limited.

To directly address the limitations of morphology we will: 1) Quantify the extent and distribution of the problem across the tree of life and across morphological modules; 2) develop methodological toolkits for reliable phylogenetic inference using morphology, with independent molecular and stratigraphic data acting as benchmarks; 3) apply those methods to important evolutionary events that rely on interpretations of morphology but have thus far proved intractable or equivocal, for example, human origins. The outcomes will address specific evolutionary hypotheses, and provide powerful tools, workflows and guides for future analyses by providing new ways of working. This project will bring morphology out of the 19th century and into the 21st.

Morphological data are essential for phylogenetic reconstruction, but are acknowledged to be intrinsically problematic. Character independence is violated due to over-saturation, and functional and developmental linkage. Furthermore, homoplasy is widespread due to convergence, and subjectivity of character definition.

This project aims to identify the scale and distribution of these problematic phylogenetic phenomena and establish new ways of working that minimize their effects. Identification of misleading morphological moieties will be achieved by 1) investigating the performance of logical a priori subdivisions and modules of morphology (e.g. cranium vs post-cranium, hard vs soft, appendage vs non-appendage, flower vs leaf), and 2) investigating the distribution and prevalence of cliques of inter-dependent and ecologically linked characters as identified a posteriori. The performance of different phylogenetic techniques will be directly compared and assessed: 1) implied weighting which minimizes homoplasy, 2) clique removal to minimize character over-saturation, 3) and Bayesian inference to overcome the problems associated with traditional parsimony analysis. Congruence with independent data will be used as a benchmark - primarily molecular trees, which are comparatively well-validated and based on abundant data. No single instance of congruence is authoritative, but meta-analysis of combined data from different sources, clades, and authors mitigates the effect of spurious congruence that might occur in any one dataset. Meta-analysis therefore provides a very powerful tool to identify legitimate broad scale patterns that are consistent for data from a range of sources, authors and clades. The application of these meta-analyses will provide an invaluable guide for future phylogenetic research.

The research findings, namely new ways of working with phylogeny, will have direct impact in a range of areas from evolution biology, to biodiversity and conservation. Much of this impact will be achieved via academic intermediaries following uptake of those guidelines, application to morphological data, and then publication and utilization of the resulting phylogenies. To maximize this, research findings will be disseminated to a broad range of phylogenetic practitioners through varied conferences and publication in high-profile and specialized journals. Outside of this, more direct impact will be achieved in the field of systematics. Studies of biodiversity and conservation are entirely reliant on taxonomy (i.e. species classification), which is in turn based on interpretation of phenotype and morphological characteristics. As such, our research findings will have a direct impact on taxonomic practices, and thus study and preservation of biodiversity. Our roadmap of aspects of morphology that better reflect evolutionary history will therefore be informative for taxonomy. We will directly target the communities of academic and non-academic taxonomists (museum professionals, curators, collectors etc.) through dedicated meetings and articles.

More broadly, the research will be a great vehicle to improve public understanding of science. As evidenced by extensive media coverage of palaeontology, the fossil record and evolution are of great interest to the public. This is especially true of human origins, which is one of the key case studies to be analysed in this project. The results from this part of the project will be prioritized for outreach, both through exhibition in collaboration with the Manchester Museum, as well as other events and press releases. Both the PI and Co-I have a strong track record of public engagement through participation in events, online presence, and broader dissemination.