Landscape drivers of genomic diversity and divergence in woodland Eucalyptus

Abstract

Spatial genetic patterns are influenced by numerous factors, and they can vary even among coexisting, closely related species due to differences in dispersal and selection. Eucalyptus (L'Héritier 1789; the “eucalypts”) are foundation tree species that provide essential habitat and modulate ecosystem services throughout Australia. Here we present a study of landscape genomic variation in two woodland eucalypt species, using whole‐genome sequencing of 388 individuals of Eucalyptus albens and Eucalyptus sideroxylon. We found exceptionally high genetic diversity (π ≈ 0.05) and low genome‐wide, interspecific differentiation (FST = 0.15) and intraspecific differentiation between localities (FST ≈ 0.01–0.02). We found no support for strong, discrete population structure, but found substantial support for isolation by geographic distance (IBD) in both species. Using generalized dissimilarity modelling, we identified additional isolation by environment (IBE). Eucalyptus albens showed moderate IBD, and environmental variables have a small but significant amount of additional predictive power (i.e. IBE). Eucalyptus sideroxylon showed much stronger IBD and moderate IBE. These results highlight the vast adaptive potential of these species and set the stage for testing evolutionary hypotheses of interspecific adaptive differentiation across environments.

Divergent lineages in a semi‐arid mallee species, Eucalyptus behriana, correspond to a major geographic break in southeastern Australia

“…have shown that E. behriana has genetic diversity, as measured by heterozygosity, consistent with what has been reported in other box eucalypts(Jordan et al, 2016;Murray et al, 2019;Supple et al, 2018) and only moderate genetic structuring between populations. Despite this, the species-wide F its distribution likely reducing gene flow between populations.…”

“…The observed support for isolation by distance in E. behriana is unsurprising given the outlying populations at extremes of the distribution, which are likely isolated from gene flow with other populations, and it does not rule out the existence of discrete population structuring. Similar patterns are seen in other widespread tree species with large population sizes across various regions of the world, including other Eucalyptus species (Bloomfield et al, 2011;Jones et al, 2013;Murray et al, 2019;Nevill, Bradbury, et al, 2014;Supple et al, 2018), Quercus (Ju et al, 2019), Populus (Keller et al, 2010), and Pinus (Potter et al, 2015), which show limited genetic structuring; with isolation by distance explaining much of the genetic divergence. In the case of eucalypts, it has been suggested that their high recombination rates (Gion et al, 2016), preferential outcrossing (Byrne, 2008;Horsley & Johnson, 2007), and substantial inbreeding depression in the case of selfing (Nickolas et al, 2019), all favor the retention of genetic diversity and a lack of geographical structuring despite lacking long distance dispersal capabilities (Booth, 2017).…”

“…ST (0.066) was the highest so far reported for any species in Eucalyptus section Adnataria (0.017-0.018 in E. albens and E. sideroxylon(Murray et al, 2019), 0.04 in E. melliodora(Supple et al, 2018), 0.02 across the E. brownii/E. populnea cline(Holman et al, 2003)), indicating greater, albeit still low, genetic differentiation between populations.…”

Scale‐dependent signatures of local adaptation in a foundation tree species

“…To filter on LD, we removed SNPs using the snpgdsLDpruning function where the correlation was >0.5 within a sliding window of 500 base-pairs, with a 500 base-pair window deemed adequate because LD decays rapidly in Eucalyptus (Murray et al, 2019). All four of these filtering steps were implemented on two sample sets; a range-wide data set, consisting of 17 widespread populations, and a reduced data set containing only samples collected with the VCH region.…”

Detecting selection in low-coverage high-throughput sequencing data using principal component analysis

“…The reduced expenses for whole genome DNA sequencing, thanks to advanced High-throughput DNA sequencing technologies, has facilitated larger sample sizes in population genetics studies in the recent years, including samples with similar genetic ancestry [ 4 , 14 , 20 , 26 , 29 , 30 ]. Identifying signatures of selection in populations of similar genetic ancestry can results in arbitrary population assignments when using methodologies that require discrete groups of populations.…”