Seminar at the Swedish Museum of Natural History on Wednesday November 25, 16:15–17:00 in the lunch/seminar room of the KÖL building, level 2:
Understanding the History of Life Using Morphology and Fossils: New Computational Approaches
Postdoctoral Associate, Yale University
Abstract. Fossils represent a unique and indispensable source of data for studying macroevolutionary processes and dynamics, as they provide us with direct glimpses of how life actually evolved on Earth. Although stratigraphic data from fossils have been used extensively in divergence time calibration, morphology remains the primary type of data that can be extracted from fossils for use in phylogenetic inference. In general, however, molecular sequence data is more widely used and trusted for phylogenetic analyses (e.g., 106,255 vs. 33,473 publications matching the topic of “molecular phylogeny” vs. “morphological phylogeny” on the Web of Science between 2010-2015). This is due both to the relative ease of generating large amounts of genetic data as well as concerns about subjectivity and the acquisition of morphological data. As a result, when phylogenetic incongruence arises between morphological/paleontological and molecular datasets, the latter is often viewed as more robust, even when the divergences being inferred occurred in deep time. Molecular data, however, are not infallible, and the meaningful influence of fossil data in phylogeny estimation and comparative analyses is well established. Therefore, efforts directed towards improving the methodological process of generating and analyzing morphological data are a priority.
In this presentation, I will discuss computational approaches that illustrate that: 1) systematic biases and misleading signal may have a profound effect on molecular phylogenetic analyses; 2) the inclusion of phenomic-scale datasets in combined analyses can affect phylogenetic inference and comparative methods, even when morphological characters are vastly outnumbered; and 3) morphological data extraction can potentially be automated and scaled up effectively and efficiently. To demonstrate these points, I use three case studies, respectively: 1) the position of turtles within the amniote tree of life; 2) the evolutionary history and origin of snakes; and 3) the evolution of shape across North Atlantic communities in planktonic foraminifera. These studies set the groundwork for future work aiming to improve computational methods for analyzing morphological and paleontological data, both in terms of data extraction and data interpretation/analysis.