Genetic basis of traits and performance in UK silver birch

Knowledge of the genomic basis of phenotypes in forest trees lags that of other economically important organisms. We sequenced the genomes of 2054 silver birch trees from 20-year-old field trials at three locations in Britain. Each trial site contains trees from the same 29 source populations, and one trial site (Drummond) is outside the climate envelope of the source sites. We discovered many single nucleotide polymorphisms (SNP), and a structural polymorphism, associated with several growth traits. Using models trained with genome-wide SNPs we estimate the genomic breeding values of 148 plus trees. We discover 117 SNPs significantly associated with source-site environmental variables, but genetic offset estimates using these SNPs provide poor predictions of tree growth rates (Drummond excluded). They are outperformed in this task by offset estimates using thousands of non-significant SNPs. Such estimates for Drummond in current climates are higher than for source sites under a future worst case climate scenario.