Application of high-throughput methods to reconstruct the evolutionary history of amherstieae by using the fossil pollen record of Striatopollis catatumbus and extant geographical records of the tribe

To have a better understanding of the evolutionary history of the different plant lineages, especially in the tropics, it is important to integrate the study of fossil records with extant occurrences. Fossil records provide information of origination, extinction, and dispersal events that are essential to understand the current biogeography, as well as the ecological and climatic preferences of extant taxa. This dissertation explores novel approaches to improve the taxonomic identification of fossil pollen. It also applies an ecological analysis to review how climate has played a role in the modern geographic distributions of tropical plants, especially the legume tribe Amherstieae. First, superresolution confocal microscopy was used to increase the morphological resolution of modern and fossil pollen grains to detect diagnostic features of the pollen structure. This method was selected because it can take images of the internal and external structure of the pollen wall similar to electron microscopy, and it can do this with less-destructive preparation techniques. As an optical microscope, it can be used to image and analyze rare grains. This last advantage is important for the analyses of pollen fossils, in which rare grains or damaged grains are unfrequently selected for high-resolution analysis. Second, optical superresolution imaging was combined with machine learning identification to create a higher throughput and quantitative workflow for palynological identifications. For this, we analyzed individual pollen grains of the tropical morphospecies Striatopollis catatumbus. With these two methods, it was possible to taxonomically separate pollen grains with visibly similar morphologies under traditional microscopy. It was also possible to constrain taxonomic identifications of individual pollen grains that were at genus level. The application of high-resolution imaging and machine learning in pollen demonstrated that fossil pollen is more diverse than previously recognized. It also showed that pollen data provides more information of the evolutionary history of plant lineages than previously estimated. The concept of niche conservatism in different groups of angiosperms is critical to understanding what environmental factors have driven their current geographic distributions. In the final chapter of this dissertation, the concept of niche conservatism was reviewed in the extant legume tribe Amherstieae using climatic variables. This tribe was selected because fossil pollen grains of Striatopollis catatumbus have been associated with several genera of this tribe. Before assessing how past climate and geological changes affected the distribution of the tribe in the fossil record, it was important first to evaluate how climate variables play a role in the current distribution of the tribe. It was also important to review if there are any environmental differences between geographical regions that indicate shifts in the ecological niches. This study demonstrates that temperature is the limiting factor in the biogeographic distribution of the tribe, while precipitation appears to correlate with variations in distribution within the members of the tribe. It also highlights that the ecological niche of the tribe has been conserved among its members through time and space, and it shows the critical importance of conserving tropical environments occupied for this tribe. In addition, it sheds light on how past changes in temperature and precipitation could affect the geographical distribution of the group across the tropics.LimitedAuthor requested closed access (OA after 2yrs) in Vireo ETD syste