Gearty, W. and Gauthier, J. Resolving the Relationships of the Squamate Tree of Life: An Assessment of New Approaches and Problems. Journal of Vertebrate Paleontology, Program and Abstracts, 2014, 136.
[SVP Oral Presentation]
Since the division of The Deep Scaly Project into separate morphological and molecular teams, a truly integrated project of wide scope has not been attempted. Much more can be done to understand how the members of Squamata are related to one another through an approach that combines the importance of both morphological and molecular evolution. Here we have developed a novel three-step methodological approach to squamate phylogenetics that incorporates the newest phylogeny-creating techniques and data from previous morphological and genetic analyses. First, we analyze a large squamate morphological dataset using Lewis's Mkv model under both a Bayesian and maximum likelihood framework. Second, we incorporate a previously constructed squamate DNA dataset and analyze the combined data within a 'total evidence' framework. Finally, we adopt a methodology that treats genes, rather than nucleotides, as the character of interest.
We find that the separate analyses of the morphological and molecular datasets, even under Bayesian and maximum likelihood frameworks, still result in drastically different relationships between higher-order clades within Squamata. Additionally, we find that the combination of these two datasets results in a phylogeny with limited support for either topology, although it definitively leans in the direction of the molecular results. Finally, by reducing the molecular dataset to gene characters, we find significantly lower support for the higher-order relationships that are strongly supported in previous analyses. By combining these data with our morphological dataset, we discover that we have inversed the effect of the power in numbers problem.
We conclude that combining datasets, although possibly detrimental to results, should be treated as a source of understanding how the datasets may differ and how they may reflect different evolutionary histories.