Search

Long-lived species of trees, especially conifers, often display weak patterns of reproductive isolation, but clear patterns of local adaptation and phenotypic divergence. Discovering the evolutionary history of these patterns is paramount to a generalized understanding of speciation for long-lived plants. We focus on two closely related yet phenotypically divergent pine species, Pinus pungens and P. rigida, that co-exist along high elevation ridgelines of the southern Appalachian Mountains. In this study, we performed historical species distribution modeling (SDM) to form hypotheses related to population size change and gene flow to be tested in a demographic inference framework. We further sought to identify drivers of divergence by associating climate and geographic variables with genetic structure within and across species boundaries. Population structure within each species was absent based on genome-wide RADseq data. Signals of admixture were present range-wide, however, and species-level genetic differences associated with precipitation seasonality and elevation. When combined with information from contemporary and historical species distribution models, these patterns are consistent with a complex evolutionary history of speciation influenced by Quaternary climate. This was confirmed using inferences based on the multidimensional site frequency spectrum, where demographic modeling inferred recurring gene flow since divergence (2.74 million years ago) and population size reductions that occurred during the last glacial period (~ 35.2 thousand years ago). This suggests that phenotypic and genomic divergence, including the evolution of divergent phenological schedules leading to partial reproductive isolation, as previously documented for these two species, can happen rapidly, even between long-lived species of pines. [ABSTRACT FROM AUTHOR]

Explaining the origin and evolutionary dynamics of the genetic architecture of adaptation is a major research goal of evolutionary genetics. Despite controversy surrounding success of the attempts to accomplish this goal, a full understanding of adaptive genetic variation necessitates knowledge about the genomic location and patterns of dispersion for the genetic components affecting fitness-related phenotypic traits. Even with advances in next-generation sequencing technologies, the production of full genome sequences for non-model species is often cost-prohibitive, especially for tree species such as pines where genome size often exceeds 20 to 30 Gbp. We address this need by constructing a dense linkage map for foxtail pine (Pinus balfouriana Grev. & Balf.), with the ultimate goal of uncovering and explaining the origin and evolutionary dynamics of adaptive genetic variation in natural populations of this forest tree species. We utilized megagametophyte arrays (n = 76-95 megagametophytes/tree) from four maternal trees in combination with double digest restriction-site associated DNA sequencing (ddRADseq) to produce a consensus linkage map covering 98.58 % of the foxtail pine genome, which was estimated to be 1276 cM in length (95 % CI, 1174 to 1378 cM). A novel bioinformatic approach using iterative rounds of marker ordering and imputation was employed to produce single-tree linkage maps (507-17,066 contigs/map; lengths, 1037.40-1572.80 cM). These linkage maps were collinear across maternal trees, with highly correlated marker orderings (Spearman’s ρ>0.95). A consensus linkage map derived from these single-tree linkage maps contained 12 linkage groups along which 20,655 contigs were non-randomly distributed across 901 unique positions (n=23 contigs/position), with an average spacing of 1.34 cM between adjacent positions. Of the 20,655 contigs positioned on the consensus linkage map, 5627 had enough sequence similarity to contigs contained within the most recent build of the loblolly pine (Pinus taeda L.) genome to identify them as putative homologues containing both genic and non-genic loci. Importantly, all 901 unique positions on the consensus linkage map had at least one contig with putative homology to loblolly pine. When combined with the other biological signals that predominate in our data (e.g., correlations of recombination fractions across single trees), we show that dense linkage maps for non-model forest tree species can be efficiently constructed using next-generation sequencing technologies. We subsequently discuss the usefulness of these maps as community-wide resources and as tools with which to test hypotheses about the genetic architecture of local adaptation. [ABSTRACT FROM AUTHOR]