This project will develop new methods for combining fossil and living species in phylogenetic analyses of evolutionary history. Information from DNA sequence data and proteins is often used to reconstruct evolutionary trees for living species. Data from the phenotype is also critically important in such analyses in order to investigate how the phenotype changes through time along the tree, and to assess how these changes might be correlated. One goal of this project will be to use morphometric measurements of shape to assess the placement of fossil species on phylogenetic trees. The project will also develop methods and analytical programs for inclusion of multiple fossils and multiple shape characters in integrative phylogenetic analyses.
The project will also investigate how morphometric methods can be used to describe the evolution of organismal shape along evolutionary trees. Morphometric methods applied to shapes are based on landmark points (points with names) on organisms. Although good methods exist for collecting the locations of these points and for describing the shapes these points make, better methods for describing how those shapes change throughout evolutionary time are required in modern studies, and will be developed here. The project will train several graduate students in phylogenetic computing and morphometric analyses, and will provide outreach and training in new methods to the scientific community.
This project studies methods for understanding the evolution of measurable characters. An important part of it studies morphometrics, which is the description of shapes and sizes. It takes a 2-dimensional or 3-dimensional form and records it by Cartesian coordinates of sets of points. Some of the points are called landmarks, such as the corner of a mouth or a place on a skull where three plates of bone meet, while others, called semilandmarks, are carefully spaced points on curves in-between landmarks (such as the suture between two bones of the skull) or on surfaces in-between semilandmark curves (such as the surfaces of the outer skull itself). Given landmarks and semilandmarks from typical forms of different species, we use an evolutionary tree of the species to understand which parts of the shapes have changed most, and which ones change in similar ways. We hope to understand what changes of shape are favored by natural selection, and which ones are simply co-occuring because the same genes affect both parts. When we do this, there is the problem of placing the specimen and of rotating it to make the coordinates of different specimens comparable. We have developed new, better statistical methods for correcting for the placements and correcting for the rotations. A computer program, Contrast, has been written to handle morphometric data this way. It will released soon in the PHYLIP package, which does statistical analysis of the fit between data and evolutionary trees. One of the major problems here is to understand how changes in size affect changes in shape. If a species gets larger, will its body become relatively thinner or relatively fatter? This is a problem called allometry. The Contrast program can now separate size and shape changes, and also can detect when the evolution of the shape is affected by the evolution of size. Another major problem comes from our awareness as biologists that natural selection operates at all times in an organism's life from conception to maturity. Patterns of evolutionary changes in size, shape, or both can apply either to the management of growth or the management of form. These changes can be either specific to individual parts or shared across parts, and, whether part-specific or integrated, they may or may not be recognizable as having anything to do with what we already understand about how those forms work as biological organisms (the concern usually called "function"). All these distinctions have to be taken seriously if we are to really understand how forms evolve along evolutionary trees. So the other part of our work under this grant has involved developing new formulas for making distinctions of interpretation like these based on the diagrams that come from the new computed analyses of real morphometric data sets. One kind of new formula deals with deciding if changes are part-specific or not; another kind looks at trends over short periods of time within a fixed branch of the evolutionary tree; a third considers ways in which our understanding of how forms work (i.e. forms with longer legs run faster, forms with bigger jaws can chew bigger prey) get rewritten so they can be tested using this combination of evolutionary and morphological data. The methods developed in this project improve on previous morphometric methods by not only taking the evolutionary tree into account, but also by more efficiently measuring which parts of the shape change fastest, and which parts change in related ways. This is an important step in developing a new approach which we call "phylomorphometrics".
