ABI INNOVATION: PHYLOGENETIC METHODS FOR IMPUTING MISSING PLANT TRAIT VALUES IN GLOBAL DATABASES

点击翻译按钮获取中文摘要

An award is made to Michigan State University to develop methods to use the known shared evolutionary history of species to estimate the trait values missing in global databases. Ecologists have demonstrated a strong linkage between the diversity of how plants function and the health and stability of our ecosystems. This linkage can therefore be utilized to generate predictive models of how vegetation responds to climate from local to global scales. However such modeling on global scales is fundamentally limited by a lack of detailed maps of the diversity of plant function on continental scales. This limitation is generally the result of large holes in global plant trait databases that will not be adequately filled through additional field campaigns in the near future. Thus a key challenge this project will address by developing derive computational tools is the estimation of missing trait values in global databases. In particular, previous work has shown that closely related species tend to have similar traits due to common descent. This evolutionary non-independence in trait data already collected can be harnessed to make reliable estimations of trait data that cannot be collected in the near future. This project tests the suitability of estimation methods on real and simulated datasets to determine when, where and why the methods are robust. The ultimate goal of this research is to identify and test a series of methods that can rapidly and reliably fill in the gaps of global trait databases to facilitate the mapping of plant function on continental scales. It is expected that such maps will then be utilized to generate more refined models of vegetation structure and dynamics in the future.The project includes training and mentoring of a postdoctoral researcher and graduate student and collaborative research with colleagues in Denmark. The project also includes phylogenetic and functional trait training workshops for graduate students and early career researchers. The research will make several previously unpublished datasets publicly available and will lead to the development and and dissemination of novel R code to impute functional trait values in global databases using large phylogenetic trees. More information on the project can be found at: https://www.msu.edu/~swensonn