PhD: An Information-Theoretic Measure of Tree Similarity: How Resilient are Methods for Reconstructing Evolutionary History?

Many of the most fundamental steps in the reconstruction of evolutionary history are underpinned by metrics of tree similarity – whether this involves evaluating the performance of competing methods of phylogenetic reconstruction, or exploring how robust a dataset is to various sources of bias or error. Yet no meaningful measure of tree similarity exists. This project will take a new approach to this problem, using the methods of information theory to quantify the amount of information that an arbitrarily large group of trees has in common.

A computationally efficient approach to calculating this value will allow a robust approach to a number of outstanding biological problems: which methods should be used to infer cladistic relationships and evolutionary history from different datasets? What do the results of different approaches have in common? Can palaeontological data contribute to our understanding of deep animal relationships, or do the processes of decay systemically and indelibly distort phylogenetic signal? How compatible are the phylogenetic signals presented by the genotype and from the phenotype, and how does this vary at different levels of the taxonomic hierarchy?