Mosaic genetic differentiation along environmental and geographic gradients indicate divergent selection in a white pine species complex

Abstract

We explored the role of isolation by environment in a white pine species complex: Pinus flexilis, Pinus strobiformis and Pinus ayacahuite distributed from Canada to Central America. We predict that species differentiation would match genetic structure of candidate genes associated with significant differences in climatic niche in the species complex. To test this prediction, we sequenced five candidate genes for drought tolerance and three housekeeping genes, in individuals from across the entire range of each species. We performed neutrality tests, estimated genetic differentiation and performed partial mantel correlations, to test for isolation by environment in the species complex. Our results show that different loci vary in degrees of genetic differentiation within species and contrast in patterns of differentiation among species. This is considered to be a mosaic pattern of genetic differentiation. There was also significant isolation by environment in candidate genes. P. flexilis was genetically differentiated for candidate genes and P. ayacahuite for housekeeping genes. There was also an overall pattern of shared ancestral polymorphism followed by independent evolution. Nonetheless, all loci together recovered groups that correspond to the recognized taxonomy. In conclusion, the pattern of isolation by environment in candidate genes support the idea of ecologically driven differentiation of this species complex, especially in the case of P. flexilis. The observed difference in housekeeping genes between P. strobiformis and P. ayacahuite can be due to limited gene Electronic supplementary material The online version of this article (flow. The mosaic pattern of differentiation suggests that speciation is recent and ecological differences could be a factor in the diversification of pines in North America.

The role of hybridization during ecological divergence of southwestern white pine (Pinus strobiformis) and limber pine (P. flexilis)

“…Thus, given sufficient time, even under a model of speciation with gene flow, such coupling effects will ensure the maintenance of species boundaries relative to the action of either factor alone (Barton & De Cara 2009). Specifically, in our focal species, previous work using candidate genes for drought stress provides evidence for divergent selection driving speciation, despite low genome-wide levels of differentiation (Moreno-Letelier & Barraclough 2015). Although a thorough examination of exome-wide variation remains to be done, the correlation of h with drought related variables when coupled with the work of Moreno-Letelier & Barraclough (2015) implies that adaptive responses to drought stress likely contributed to the origin and maintenance of species boundaries in this system.…”

“…The lack of islands of divergence is expected in conifers, given the prevalence of polygenic architectures defining continuous trait variation across species boundaries and the expected prevalence of soft sweeps (Pritchard & Rienzo 2010; Alberto et al 2013; Rajora et al 2016; Lind et al 2017). Alternatively, given the large and complex genomes of conifers (reviewed by De La Torre et al 2014a) our ddRADseq markers underrepresented genic regions, which are often identified as islands of divergence (Nosil & Feder 2012; Zhou et al 2014; Moreno-Letelier & Barraclough 2015; Marques et al 2017). For example, Moreno-Letelier & Barraclough (2015) demonstrated the potential for islands of divergence at drought-associated genes, which had a high average F ST value of 0.33 (0.09–0.4) compared to the genome-wide estimate from this study ( F ST-species = 0.02).…”

“…Alternatively, given the large and complex genomes of conifers (reviewed by De La Torre et al 2014a) our ddRADseq markers underrepresented genic regions, which are often identified as islands of divergence (Nosil & Feder 2012; Zhou et al 2014; Moreno-Letelier & Barraclough 2015; Marques et al 2017). For example, Moreno-Letelier & Barraclough (2015) demonstrated the potential for islands of divergence at drought-associated genes, which had a high average F ST value of 0.33 (0.09–0.4) compared to the genome-wide estimate from this study ( F ST-species = 0.02). Future investigations using candidate gene approaches or exome capture might thus be able to identify islands of divergence in conifers, although evidence of adaptation in complex genomes often also appears within intergenic regions (Li et al 2012).…”

“…Our focal species inhabit a wide latitudinal range in the western part of North America, with a putative area of sympatry located in the southern Rocky Mountains and Colorado Plateau in which morphological evidence points towards the occurrence of hybridization (Steinhoff & Andresen 1971; Tomback & Achuff 2010). These species also display limited differences in morphological and reproductive traits (Benkman et al 1984; Tomback et al 2011; Bisbee 2014) and show evidence of local adaptation to the heterogeneous climatic conditions across their geographical range and also within the area of putative hybridization (Steinhoff & Andresen 1971; Moreno-Letelier et al 2013; Borgman et al 2015; Moreno-Letelier & Barraclough 2015; Goodrich et al 2016). To examine the processes influencing species boundaries between these two conifer species, we asked three questions: (1) Is there niche divergence among P. strobiformis , P. flexilis and the putative hybrid zone?…”

The role of hybridization during ecological divergence of southwestern white pine (Pinus strobiformis) and limber pine (P. flexilis)

“…Islands of divergence are often expected only under certain genetic architectures and selection scenarios which have been shown to be less prevalent in conifers (Alberto et al., ; Lind et al., ; Pritchard & Di Rienzo, ; Rajora, Eckert, & Zinck, ). Alternatively, given the large and complex genomes of conifers (reviewed by De La Torre, Birol et al., ), our ddRADseq markers likely underrepresented genic regions, which are often identified as islands of divergence (Marques et al., ; Moreno‐Letelier & Barraclough, ; Nosil & Feder, ; Zhou, Zhang, Liu, Wu, & Savolainen, ). For example, Moreno‐Letelier and Barraclough () demonstrated the potential for islands of divergence at drought‐associated genes, which had a high average F ST of 0.33 (0.09–0.40) as compared to the genome‐wide estimate from this study ( F ST‐species = 0.02).…”