Strong divergent selection at multiple loci in two closely related species of ragworts adapted to high and low elevations on Mount Etna

Abstract

Defining what constitutes a species and understanding how new species form are central questions in evolutionary biology. The idea that speciation is driven by adaptation to distinct environmental conditions has a long history going back to Darwin's concept of new species formation by means of natural selection. Although natural selection is the main driver of evolutionary changes, its role in creating barriers to gene flow is not well understood. Adaptation to distinct environments (divergent selection) may lead to differentiation in loci responsible for adaptation or even fixation of different alleles in isolated populations. This can generate a mosaic with different genomic AbstractRecently diverged species present particularly informative systems for studying speciation and maintenance of genetic divergence in the face of gene flow. We investigated speciation in two closely related Senecio species, S. aethnensis and S. chrysanthemifolius, which grow at high and low elevations, respectively, on Mount Etna, Sicily and form a hybrid zone at intermediate elevations. We used a newly generated genome-wide single nucleotide polymorphism (SNP) dataset from 192 individuals collected over 18 localities along an elevational gradient to reconstruct the likely history of speciation, identify highly differentiated SNPs, and estimate the strength of divergent selection.We found that speciation in this system involved heterogeneous and bidirectional gene flow along the genome, and species experienced marked population size changes in the past. Furthermore, we identified highly-differentiated SNPs between the species, some of which are located in genes potentially involved in ecological differences between species (such as photosynthesis and UV response). We analysed the shape of these SNPs' allele frequency clines along the elevational gradient. These clines show significantly variable coincidence and concordance, indicative of the presence of multifarious selective forces. Selection against hybrids is estimated to be very strong (0.16-0.78) and one of the highest reported in literature. The combination of strong cumulative selection across the genome and previously identified intrinsic incompatibilities probably work together to maintain the genetic and phenotypic differentiation between these species -pointing to the importance of considering both intrinsic and extrinsic factors when studying divergence and speciation.

Rapid homoploid hybrid speciation in British gardens: The origin of Oxford ragwort (Senecio squalidus)

“…Conversely, the hybrid individuals collected on Mt Etna showed variable degrees of admixture between the two parental species, and based on the admixture proportions none appeared to be a first‐generation hybrid (Figure 2a). This is in agreement with previous results using different genetic markers (James & Abbott, 2005; Wong et al., 2020) and suggests that F1 hybrids on Mt Etna are rare. The stronger clustering of all hybrid individuals with S. chrysanthemifolius accords with the relatively low elevation where hybrid seeds were collected for this study (1,353–1,530 m; Table ) which is closer to the distributional limits of typical S. chrysanthemifolius (below 900 m) than S. aethnensis (above 2,500 m; Abbott & Brennan, 2014), and is more likely to reflect the geographic origin of the putative hybrid material introduced to the Oxford Botanic Garden in the late‐17th century (Harris, 2002).…”

“…using different genetic markers (James & Abbott, 2005;Wong et al, 2020) and suggests that F1 hybrids on Mt Etna are rare. The stronger clustering of all hybrid individuals with S. chrysanthemifolius accords with the relatively low elevation where hybrid seeds were collected for this study (1,353-1,530 m; Table S1) which is closer to the distributional limits of typical S. chrysanthemifolius (below 900 m) than S. aethnensis (above 2,500 m; , and is more likely to reflect the geographic origin of the putative hybrid material introduced to the Oxford Botanic Garden in the late-17th century (Harris, 2002).…”

“…Colonisation of the UK by S. squalidus after escaping cultivation at the Oxford Botanic Garden has been extensively recorded by British botanists to give an extremely accurate chronological record (reviewed in Abbott et al, 2009), but the early evolutionary history of this species remains elusive (Abbott et al, 2009;Harris, 2002). (James & Abbott, 2005;Wong et al, 2020). Conversely, Model 2 assumes a very recent hybridisation event between S. aethnensis and S. chrysanthemifolius as giving origin to S. squalidus.…”

Genetic, morphological, and niche variation in the widely hybridizingRhus integrifolia‐Rhus ovataspecies complex

“…Widespread pollen flow (both historical and contemporary), potential environmental barriers to external propagule recruitment in established populations, and differences in plastid versus nuclear DNA effective population sizes may all contribute to the discrepancy in population structure among localities. Genome‐scale data would allow several hypotheses to be tested regarding demographic history, patterns of gene flow between the two species and their introgressants, the adaptive value (if any) of introgression, and detailed analysis of maintenance of hybrid zones between the two species in California (e.g., Wong et al, 2020). For example, a comparison of which regions of the genome have experienced higher or lower rates of introgression would be particularly insightful, especially in the context of putatively adaptive or maladaptive introgression (e.g., Grant & Grant, 1998; Rieseberg & Burke, 2001).…”

“…The acquisition of genome‐scale variation and implementation of hybrid zone analyses (sensu Abbott, 2017; Ma et al, 2019; Wong et al, 2020) should be informative in determining patterns of gene flow among these two species (Taylor & Larson, 2019). Such data will allow the detection of adaptive variants, and regions of the genome that experience higher or lower levels of gene flow than regions under neutral expectations (Whitney, Ahern, Campbell, Albert, & King, 2010).…”