Genetic evidence for species cohesion, substructure and hybrids in spruce

Abstract

The origin and history of species are shaped by various evolutionary dynamics, including their persistence in the face of potential gene flow from related taxa. In this study, we use broad geographical and taxonomic sampling (2,219 individuals) to establish the distribution of species, hybrids and cryptic genetic variation within the conifer genus Picea (spruce) across western North America. We demonstrate that the six species of spruce in this region are distinguishable based on their genetic composition, and that the more closely related Engelmann spruce (P. engelmannii) and white spruce (P. glauca) have generated numerous and widespread hybrids. These hybrids occur in the central Rocky Mountains, well to the south of the well‐established region of admixture in Canada. Additionally, we provide evidence for subdivision within Engelmann spruce, manifested as a southern Rocky Mountains form, and a northern Rocky Mountain and Cascade mountains (western) form. In the intervening central Rocky Mountains region (forests in Wyoming and adjacent states) we found primarily individuals with admixed ancestry. Following their origin, these species of spruce have interacted repeatedly and in different geographical contexts. Multiple pairs of species have been shown to hybridize, yet the species persist and retain distinguishable compositions. At the same time, large geographical areas exist where hybrids are pervasive. Consequently, spruce provide a case study for the maintenance of species boundaries, particularly for how widespread hybridization need not lead to the collapse and loss of species.

Host plant associations and geography interact to shape diversification in a specialist insect herbivore

“…We sampled B. treatae from 58 sites ( N = 14 Qf , N = 36 Qv, and N = 8 Qg ) distributed throughout the geographic ranges of the host plants (Figure ; Table ; Appendix ). We designed sampling to minimize apparent population genetic structure that can arise by gaps in the distribution and density of sample locations (Haselhorst, Parchman, & Buerkle, ; Lawson, van Dorp, & Falush, ). Belonocnema treatae exhibits heterogony in which temporally distinct sexual and asexual generations alternate to complete the life cycle (Lund et al, ).…”

“…We sampled B. treatae from 58 sites (N = 14 Qf,N = 36 Qv,and N = 8 Qg) distributed throughout the geographic ranges of the host plants ( Figure 1; Table S1; Appendix S1). We designed sampling to minimize apparent population genetic structure that can arise by gaps in the distribution and density of sample locations (Haselhorst, Parchman, & Buerkle, 2019;Lawson, van Dorp, & Falush, 2018 (Gokhman, Ott, & Egan, 2015) and because the single-chambered leaf galls are easily sampled in large numbers to maximize genetic variation. We collected leaf galls prior to adult emergence during October and November in both 2015 and 2016.…”

Model‐based genotype and ancestry estimation for potential hybrids with mixed‐ploidy

“…This statistical approach yields a probability of assignment of individuals to demes (the K-means clusters), without admixture. We have used the estimated assignment probabilities as mean initialization values (with some variance) for q in the entropy model and software (e.g., Gompert, Lucas, et al, 2014;Haselhorst et al, 2019;Mandeville et al, 2015). Additionally, starting values for the admixture proportions could come from taxonomic labels or justified strata in the sampling.…”

Model-based genotype and ancestry estimation for potential hybrids with mixed-ploidy

“…Given the potentially large number of loci and individuals in a contemporary study, the values (with some variance) for q in the entropy model and software (e.g., Gompert et al, 2014b;Mandeville et al, 2015;Haselhorst et al, 2019). Additionally, starting values for the admixture proportions could come from taxonomic labels or justified strata in the sampling.…”

“…This statistical approach yields a probability of assignment of individuals to demes (the K-means clusters), without admixture. We have used the estimated assignment probabilities as mean initialization values (with some variance) for q in the entropy model and software (e.g.,Gompert et al, 2014b;Mandeville et al, 2015;Haselhorst et al, 2019). Additionally, starting values for the admixture proportions could come from taxonomic labels or justified strata in the sampling.The software implementation uses the initial q values to compute the initial population allele frequency in each of the K populations.…”