William J. Plumb

Societal Impact Statement: The world‐wide diversity of ash trees includes genetic information encoding resistance to the ash dieback fungus and the emerald ash borer beetle, which are currently devastating ash populations in Europe and North America. In order to mobilise this genetic diversity in conventional breeding programmes, we need to be able to accurately identify ash species from around the world and cross them with one another. Here, we present a preliminary genetic barcoding system for ash, and a series of hybridisation experiments between European ash and other species. Two of the hybrids show early promise against ash dieback. Summary: Native ash tree species in Europe and North America are being devastated by ash dieback and the emerald ash borer, respectively. As worldwide ash species differ in their level of susceptibility to these threats, hybrid breeding may allow resistance to be transferred among species. However, we do not know the extent to which distantly related ash species can be crossed, and many ash species are difficult to identify from morphology alone leading to some mislabelling in living collections. Here, we initiate development of a genetic barcode system for the identification of Fraxinus species based on low‐copy‐number protein coding genes. We also conduct experimental crosses among ash species in different sections. Three genes are effective in identifying ash samples to sectional level but only in some cases to species level. They highlight that Fraxinus mandshurica , Fraxinus platypoda and Fraxinus chiisanensis may be frequently mistaken for one another in living collections. We succeeded in generating 10 wide hybrid plants: two of section Melioides (species: Fraxinus pennsylvanica ) × section Fraxinus (species: Fraxinus excelsior ) and eight of section Ornus (species unclear) × section Fraxinus (species: F. excelsior ). One hybrid from each of our crosses has survived natural infection with the ash dieback pathogen in Ireland. We also discovered a hybrid between section Melioides (species: F. latifolia ) × section Fraxinus (species: F. excelsior ) formed spontaneously in the ash collection at Kew. Our findings facilitate the utilisation of global ash species diversity in response to novel threats and highlight the difficulty of designing a barcoding system capable of distinguishing all species of a genus.

Native ash tree species in Europe and North America are being devastated by ash dieback and the emerald ash borer, respectively. As worldwide ash species differ in their level of susceptibility to these threats, hybrid breeding may allow resistance to be transferred among species. However, we do not know the extent to which distantly related ash species can be crossed, and many ash species are difficult to identify from morphology alone leading to some mislabelling in living collections. Here, we develop a genetic barcode system for the identification of Fraxinus species based on three low-copy-number protein coding genes. We also conduct experimental crosses among ash species in different sections. Our barcodes are effective in identifying ash samples to sectional level and in some cases to species level, and can also identify hybrids. They highlight that F. mandshurica , F. platypoda and F. chiisanensis may be frequently mistaken for one another in living collections. We succeeded in generating ten wide hybrid plants: two of section Melioides (species: F. pennsylvanica ) ⨉ section Fraxinus (species: F. excelsior ) and eight of section Ornus (species unclear) ⨉ section Fraxinus (species: F. excelsior ). One hybrid from each of our crosses has survived natural infection with the ash dieback pathogen in Ireland. We also discovered a hybrid between section Melioides (species: F. latifolia ) ⨉ section Fraxinus (species: F. excelsior ) formed spontaneously in the ash collection at Kew. Our findings facilitate the deployment of global ash species diversity in response to alien pests and pathogens.

Rapid evolution via small shifts in allele frequencies at thousands of loci are a long- standing neo-Darwinian prediction but are hard to characterize in the wild. European ash tree ( Fraxinus excelsior ) populations have recently come under strong selection by the invasive fungal pathogen Hymenoscyphus fraxineus . Using genomic prediction models based on field trial phenotypes and 7,985 loci, we show a shift in genomically estimated breeding values in an ancient woodland, between adult trees established before the epidemic started and juvenile trees established since. Using simulations, we estimate that natural selection has eliminated 31% of the juvenile population. Thus, we document a highly polygenic heritable micro-evolutionary adaptive change over a single generation in the wild. One-Sentence Summary: Subtle changes at thousands of genomic locations in one generation allow woodland trees to respond to a new selective pressure

Societal Impact Statement: The largest populations of veteran oak trees in Europe are found in British parklands: managed wood pastures up to 1000 years old. Here, we present genomic evidence that parkland oak populations harbour considerable diversity and grew from local seed sources. We found some evidence for natural regeneration of offspring and for hybridization between pedunculate and sessile oak. We detected signatures of past gene flow between these two species and few regions of high differentiation within their genomes. Future expansion of this dataset may allow us to test for a genomic basis of acute oak decline, a syndrome of particular concern in parkland environments.Summary: The two predominant oak species in Europe, Quercus robur (English or pedunculate oak) and Quercus petraea (sessile oak), have a long history of human intervention. In Britain, a common management regime is parklands, often surrounding country houses. Little is known about how seeds were sourced for parkland oak populations nor the genetic diversity that they contain. We sequenced the whole genomes of 386 trees from four British parkland sites to characterize the nuclear and chloroplast genetic structure and diversity of oak parkland populations. We assembled the chloroplast genomes and matched these to restriction enzyme fragment chloroplast haplotypes found in previous studies of ancient woodlands. We found over two million high‐quality nuclear single‐nucleotide polymorphisms (SNPs), allowing us to identify 360 Q. robur, 10 Q. petraea and 16 hybrid individuals with confidence. We identified 81 coding regions exhibiting strong differentiation between the two species. We found evidence for selective sweeps in Q. robur near some regions containing genes with putative involvement in stress tolerance. We detected a few very close relatives within some sites, suggesting natural regeneration or local seed planting. There was little differentiation among the Q. robur populations at the four sites. Chloroplast genomes found in each parkland tended to be similar to those of local ancient woodlands, suggesting that they were derived from local seed sources.

The two predominant oak species in Britain are Quercus robur (English or pedunculate oak) and Q. petraea (sessile oak). We sequenced the whole genomes of 386 oak trees from four British parkland sites and found over 50 million nuclear single nucleotide polymorphisms (SNPs), allowing us to identify 360 Q. robur , ten Q. petraea and 16 hybrid individuals using clustering methods. Comparing Q. robur and Q. petraea trees from Attingham Park, we found that the nuclear genomes of the two species are largely undifferentiated but identified 81 coding regions exhibiting strong interspecific differentiation. The nuclear genomes of our 360 Q. robur individuals showed no clear differentiation among the four parkland sites. Scans for selective sweeps in Q. robur highlighted regions containing genes with putative involvement in stress tolerance, one of which was moderately differentiated from Q. petraea . Reconstructions of past effective population sizes suggested a long population size decline in both Q. robur and Q. petraea over the Pleistocene, but population growth after the last glacial maximum. We assembled the whole chloroplast genomes of 287 Q. robur , 8 Q. petraea and 14 hybrid trees. In a phylogenetic network, these fell into five major haplotypes, which were shared among species but differed in frequency among parkland sites. We matched our chloroplast genome haplotypes to restriction enzyme fragment haplotypes identified in older studies that had surveyed ancient woodlands in Britain and much of Europe. This suggested that the parkland populations in our study derive from local seed sources.

Genome-wide discovery of candidate genes for functional traits within a species typically involves the sequencing of large samples of phenotyped individuals ¹ , or linkage analysis through multiple generations ² . When a trait occurs repeatedly among phylogenetically independent lineages within a genus, a more efficient approach may be to identify genes via detection of amino acid residues shared by species possessing that trait ^3,4 . Here, by taking this approach, we identify candidate loci in the genus Fraxinus (ash trees) for resistance to the emerald ash borer beetle (EAB; Agrilus planipennis ), a pest species that appears innocuous to otherwise healthy ash in its native East Asian range ⁵ but is highly destructive in North America ⁶ and poses a threat to ash trees in Europe ⁷ . Assembling whole genome sequences for 24 diploid species and subspecies of ash, and estimating resistance to EAB for 26 taxa from egg bioassays, we find 53 genes containing amino acid variants shared between two or more independent Fraxinus lineages with EAB-resistant species, that are unlikely to be due to chance or undetected paralogy. Of these, seven genes have putative roles relating to the phenylpropanoid biosynthesis pathway and 17 are potentially connected to herbivore recognition, defence signalling or programmed cell death. We also find that possible loss-of-function mutations among our 53 candidate genes are more frequent in susceptible species, than in resistant ones. Patterns of polymorphism for the EAB-associated amino acid variants in ash trees representing different European populations suggest that selection may be able to enhance their resistance to EAB.

Societal Impact Statement: The current ash dieback epidemic in Europe caused by Hymenoscyphus fraxineus poses a key question to policy makers: whether or not to commit time and resources to the initiation of a breeding programme for the development of more resistant ash, as a long‐term policy of adaptation to the epidemic. Here we review current evidence on the potential viability of such a programme, from a biological perspective. We conclude that a breeding programme for ash aimed at resistance to current strains of H. fraxineus in the British Isles is biologically feasible.Summary: To evaluate the viability and feasibility of a future breeding programme to produce trees resistant to an emerging pest or pathogen, it is helpful to ask the following questions: How much variation in resistance exists in tree populations? To what extent is this resistance heritable? How many genetic loci are involved? What level of resistance is found in other species of the same genus? Here, we survey current knowledge of these issues in relation to the degree of resistance of European ash (Fraxinus excelsior) to H. fraxineus, the fungus causing ash dieback (ADB). Several studies have found a low frequency of heritable resistance in F. excelsior populations, which seems to be determined by many genetic loci. This suggests that a breeding programme is viable and that natural selection may also increase the mean resistance of populations over time. More research is needed on the genetic basis of resistance to ADB to understand how quickly natural selection can operate in woodlands and what acceleration may be possible in breeding programmes, including via use of genetic markers. Hybrid breeding programmes may also be a possibility, as some ash species appear to be more resistant to ADB than is F. excelsior, but more research is needed on this issue. We do not yet know if it will be possible to breed F. excelsior to have resistance to both ADB and the emerging threat of emerald ash borer. We recommend short‐term mitigation measures for the ADB epidemic and future research directions.