COS 219-3 - What dictates the genomic structure of an invasive plant with global distribution?

Abstract: The understanding of the genetic bases and modes of adaptation underlying biogeographical distribution patterns is essential to accurately estimate responses of plants to global change. High throughput data facilitate landscape genomic approaches to infer how dispersal is controlled by adaptive (e.g., isolation by environment) vs. non-adaptive drivers of gene flow (e.g., isolation by distance). Invasive plant species are suitable study models because non-native populations are often prone to rapid genomic changes as a result of colonizing a novel range. However, many studies on invasive plants compared only few and haphazardly chosen populations from large distributions, and thus, underestimated variation within each range restricting their ability to predict the native source region of non-native populations and to identify drivers of gene flow in either range.

We explore the genomic structure of 109 native 183 non-native widely distributed populations Conyza canadensis across the Northern hemisphere using ddRADseq. C. canadensis is a particularly interesting model to investigate global gene dispersal because it is a successful invader with a cosmopolitan distribution. C. canadensis is also an economically significant agricultural weed, and as such, an important model species for weed science in general. With the ddRAD data we will test the hypothesis that an interplay between environmental gradients and non-adaptive, demographic processes dictates the global genomic structure of C. canadensis.

Preliminary results suggest overall strong differentiation among populations and low variation within populations. In contrast to recent microsatellite data, we found pronounced differentiation between the genomic structures of native vs. non-native populations. Partial Mantel tests showed a significant correlation between genetic distances and both spatial and climatic distances of the populations. However, isolation by environment was more pronounced in the native ranges whereas isolation by distance was predominant in the non-native ranges.

Our preliminary results suggest that high selfing rates determine the global genomic structure of C. canadensis. Gene flow among native populations is controlled by functional connectivity whereas non-deterministic dispersal probabilities drive genetic differentiation in the non-native ranges. Upcoming analyses will be presented at ESA including clustering and least-cost corridor analyses to shed more light on how population history dictates biotic interactions across large spatio-environmental scales. We will also investigate how neutral genetic relatedness affects phenotypic differentiation under common garden conditions (pedigree mixed models), and whether genomic regions associate with phenotype (redundancy analyses and latent factor mixed models).