Development of spatially balanced collection schemes to better detect adaptive genetic variation in woody species

Better identification of adaptive genetic variation across populations of woody species would accelerate breeding efforts by identifying germplasm with a greater likelihood of tolerance to abiotic stresses, minimizing the need for lengthy, initial screenings. For these studies to yield meaningful conclusions, the accurate selection of wild germplasm is imperative. Using North American deciduous azaleas (Rhododendron sect. Pentanthera) as a system of interest, we present a methodological framework to identify suitable collection sites that maximize sampling the spatial variation for environmental variables of interest from simulated data. Sampling from simulated populations in a spatially balanced scheme, where sampling sites are spatially randomized and proportional to environmental variability, provided mixed results in the detection of preset genotype-environment associations. We accomplish this using simulated SNP marker data for 100 simulated Rhododendron sect. Pentanthera populations and a spatial generalized linear mixed model method to identify loci associated with environmental variables.