Genomic selection and genome-wide association studies to dissect quantitative traits in forest trees

Abstract: The convergence of quantitative genetics of complex traits with genomic technologies is quickly becoming an innovative approach to explore fundamental genetic questions and also have practical consequences for implementations in tree breeding. In this thesis, I used genomic selection and genome-wide association studies (GWAS) to dissect the genetic basis of quantitative traits, i.e. growth, phenology and wood property traits. I also assessed the importance of dominance and epistatic effects in hybrid Eucalyptus. Both dominance and epistasis are important in hybrids, as they are the likely contributing to the genetic basis of heterosis. To successfully implement genomic selection models, several important factors have to be considered. I found that for a good model establishment, both the size and composition of the training population, as well as the number of SNPs to be important considered. Based on the optimal models, additive, dominance and epistasis genetic effects of growth and wood traits have been estimated to evaluate genetic parameters and how these influence the prediction accuracy, which can be used in selecting elite breeding individuals or clones. I also addressed the advantage of genotyping-based analyses by showing that we could accurately correct pedigree information errors. More importantly, genotyping-based analyses capture both Mendelian segregation variation within full-sib families and cryptic genetic links through unknown common ancestors, which are not available from traditional pedigree data. GWAS were used to analyse growth and phenology related traits. Using a single-trait GWAS method, we identified a region strongly associated with the timing of bud set in Populus tremula, a trait with high heritability. For the growth related traits, we found that a multi-traits GWAS approach was more powerful than single-trait analyses as it identified more associated SNPs in hybrid Eucalyptus. Moreover, many more novel associated SNPs were identified from considering over-dominance effects in the GWAS analyses. After annotating the associated SNPs I show that these functional candidate genes were related to growth and responding to abiotic and biotic stress. In summary, the results of genomic selection and GWAS provided a deeper understanding of the genetic backgrounds of quantitative traits in forest trees.