Genome-wide analysis of allele frequency change in sunflower crop-wild hybrid populations evolving under natural conditions

Authors: CORBI, JONATHAN - University Of Georgia; BAACK, ERIC - Luther College; DECHAINE, JENNIFER - Central Washington University; Seiler, Gerald; BURKE, JOHN - University Of Georgia

Submitted to: Molecular Ecology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/25/2017
Publication Date: 2/2/2018
Citation: Corbi, J., Baack, E.J., Dechaine, J.M., Seiler, G., Burke, J.M. 2018. Genome-wide analysis of allele frequency change in sunflower crop-wild hybrid populations evolving under natural conditions. Molecular Ecology. 27(1):233-247. https://doi.org/10.1111/mec.14202.
DOI: https://doi.org/10.1111/mec.14202

Interpretive Summary: Cultivated plants and their wild relatives often hybridize due to overlapping geographical distributions and phenologies. Introgressed crop-derived alleles have been identified in at least one wild or weedy relative of 17 major cultivated crops. Potential environmental consequences of crop-wild introgression include the reduction of genetic diversity in wild populations and the evolution of increasingly weedy or invasive species. In fact, crop-wild gene flow has affected weed evolution in multiple taxa and has, in some cases, increased the fitness of weeds. Studying crop-derived alleles in wild populations is also relevant to assessing and mitigating the environmental risks associated with transgene escape. Predicting which crop-derived alleles are likely to introgress will thus require an understanding of their fitness effects under wild ecological conditions and in their genomic context. To date, crop-wild hybridization in sunflower has generally been examined via short-term studies of a few traits or genetic markers. Little is known about patterns of selection on crop-derived alleles from one generation to the next, particularly at genome-wide scales. In the current study, sunflower crop × wild hybrid recombinant inbred lines were allowed to outcross and evolve under natural ecological conditions for two generations at two field locations. Allele frequencies at a genome-wide collection of 278 single nucleotide polymorphism markers were determined for each generation. Previously published quantitative trait loci for traits that are relevant to domestication and transgenic manipulation were aligned to the SNP map, and a common garden experiment was conducted to compare these traits in third versus first generation hybrids. The current study showed that, overall, selection overwhelmingly favors wild alleles and phenotypes. However, crop alleles in certain genomic regions increased significantly in frequency, and these changes often occurred in parallel across locations. The many crop markers experiencing strong selection in both environments in the study suggest that many agronomic traits have severe fitness costs in the wild environment.

Technical Abstract: Hybridization is known to occur between cultivated and wild populations of numerous plant species. This represents a major mechanism by which a wild population’s genetic structure and evolutionary dynamics could be altered. Studying crop-derived alleles in wild populations is also relevant to assessing and mitigating the environmental risks associated with transgene escape. To date, crop-wild hybridization has generally been examined via short-term studies of a few traits or genetic markers. Little is known about patterns of selection on crop-derived alleles from one generation to the next, particularly at genome-wide scales. Here, a starting population of sunflower crop × wild hybrid recombinant inbred lines (RILs) was allowed to outcross and evolve under natural ecological conditions for two generations at two field locations. Allele frequencies at a genome-wide collection of 278 SNP markers were determined for each generation. Previously published quantitative trait loci (QTL) for traits that are relevant to domestication and transgenic manipulation were aligned to the SNP map, and a common garden experiment was conducted to compare these traits in third versus first generation hybrids. Our results show that, overall, selection overwhelmingly favors wild alleles and phenotypes. However, crop alleles in certain genomic regions increased significantly in frequency, and these changes often occurred in parallel across locations. We discuss these results in the context of potential introgression of crop alleles into wild populations, phenotypic-genotypic congruence and implications for GM crops.