Arguably the most dramatic mutation possible (whole-genome duplication, or polyploidy) results in an individual with double the chromosome number of its parents. Polyploidy is a common and ongoing phenomenon in plants, where many crops, for example, are polyploid. This project aims to examine the evolutionary potential and fate of polyploid lineages in ferns¬¬a plant group where nearly all species have experienced one or more episodes of polyploidy in their evolutionary past. The globally distributed fern genus Cystopteris encompasses a complex mix of diploid and polyploid species, making it ideal for investigating the effects of polyploidy on diversification rates. Using DNA sequence data from 200 Cystopteris individuals, each of known ploidy level, evolutionary relationships will be inferred, ancestral ploidy levels will be reconstructed, and speciation and extinction rates will be estimated through time.
Despite its ubiquity and scientific importance, the long-term potential and evolutionary dynamics of polyploidy remain contentious--are polyploid lineages "dead-ends" that are prone to rapid extinction, or do they instead drive evolutionary novelty? This empirical study is well positioned to make a key contribution to understanding the evolutionary importance of polyploidy, with implications for evolutionary theory, biodiversity conservation, and crop improvement.
Carl Rothfels has completed research for his dissertation "Phylogenetics of Cystopteridaceae: Reticulation and Divergence in a Cosmopolitan Fern Family" supported by the NSF Doctoral Dissertation Improvement Grant. This work focused on understanding how a family of ferns (the Fragile Fern and Oak Fern family—Cystopteridaceae) evolved, and especially the evolutionary role played by the combination of hybridization and genome-doubling ("allopolyploidy"). The first step in the project was to use DNA sequences from members of the Cystopteridaceae to understand how these species relate to each other, and to other ferns. These data demonstrated that members of the family form a well-supported group of relatives (a clade), and that their ancestor diverged approximately 100 million years ago from a lineage that subsequently diversified to form the bulk of the "eupolypods II"—a very diverse group of ferns that contains nearly a third of the fern species alive today. This deep evolutionary history makes the Cystopteridaceae a very important group for understanding patterns of fern evolution. Based on these results, Carl and coauthors proposed a new family-level classification for the whole eupolypods II group. Within Cystopteridaceae, this project demonstrated that there are three main genera (Acystopteris, Cystopteris, and Gymnocarpium), each morphologically and evolutionarily distinct. The family is globally distributed, locally common, and has evolved extensively by hybridization and genome duplication (polyploidy). However, the relationships among individual species (and how many species there are to begin with) are still uncertain. To get a better understanding of the number of species in Cystopteridaceae, and of the role of hybrid polyploidy (allopolyploidy) in the evolution of the family, Carl next sequenced DNA from the nucleus of the fern’s cells. Such DNA is present in multiple copies (i.e., one copy from the mother and one from the father) so can be used to infer hybridization. These results showed that allopolyploidy is rampant throughout the family, and there are many more species than previously recognized. In addition, these data provided an important insight into how the process of speciation may differ in ferns versus flowering plants, by demonstrating that a member of the Cystopteridaceae is the broadest hybrid yet known—its parents diverged from each other approximately 60 million years before the hybridization event. Normally populations lose the ability to produce viable hybrids much more quickly than this. If ferns do tend to lose this "reproductive compatibility" more slowly than other groups, it might explain why there are fewer species of ferns than, say flowering plants: it takes longer for fern populations to form new species. To be better able to infer hybridization in ferns, Carl and colleagues developed primers that allow them (and anyone else) to sequence many more regions of the fern nucleus than was previously possible. With their faster rates of evolution, multiple independent linkage groups, and biparental inheritance, these nuclear data permit a much wider range of inquiry than do the previously relied-upon chloroplast data. Using these regions (and additional data), the team inferred the first broad fern tree-of-life to be inferred from multiple nuclear markers. This evolutionary tree corroborates some conclusions earlier based entirely on chloroplast markers, and re-writes other areas of the fern tree-of-life. In particular, it strongly supports a different relationship among some of the deepest divergences in vascular plants, involving the position of the horsetail lineage. Finally, to be able to use these new resources in an economically way, Carl and colleagues developed a method that uses new sequencing technology to generate sequences from a lot of polyploids, for multiple regions, all at once. This method allows researchers, for the first time, to investigate the evolution of groups with lots of hybridization and genome doubling (which are both very common phenomena in plants) in an efficient and affordable way. Much of the training focus of this project has been on an undergraduate, Anne Johnson. For this, her first research experience, we mentored her through all stages of the scientific process. In parallel with direct work on this project, she was responsible for taking the lead in study design, data generation (including morphological data, and plastid and low-copy nuclear sequence data), analysis, and manuscript preparation and submission, which resulted in her first scientific publication.
