A genome-guided strategy for climate resilience in American chestnut restoration populations

American chestnut (<i>Castanea dentata</i>) is a deciduous tree species of eastern North America that was decimated by the introduction of the chestnut blight fungus (<i>Cryphonectria parasitica</i>) in the early 20th century. Although millions of American chestnuts survive as root collar sprouts, these trees rarely reproduce. Thus, the species is considered functionally extinct. American chestnuts with improved blight resistance have been developed through interspecific hybridization followed by conspecific backcrossing, and by genetic engineering. Incorporating adaptive genomic diversity into these backcross families and transgenic lines is important for restoring the species across broad climatic gradients. To develop sampling recommendations for ex situ conservation of wild adaptive genetic variation, we coupled whole-genome resequencing of 384 stump sprouts with genotype-environment association analyses and found that the species range can be subdivided into three seed zones characterized by relatively homogeneous adaptive allele frequencies. We estimated that 21 to 29 trees per seed zone will need to be conserved to capture most extant adaptive diversity. We also resequenced the genomes of 269 backcross trees to understand the extent to which the breeding program has already captured wild adaptive diversity, and to estimate optimal reintroduction sites for specific families on the basis of their adaptive portfolio and future climate projections. Taken together, these results inform the development of an ex situ germplasm conservation and breeding plan to target blight-resistant breeding populations to specific environments and provides a blueprint for developing restoration plans for other imperiled tree species.