Author: Randi Allyn Famula

Publisher:

Published: 2017

Total Pages:

ISBN-13: 9780355151220

Persian Walnut (Juglans regia L.) is the major nut producing species of the family Juglandaceae and is almost exclusively grown in California, with a farm gate value of $1.8 Billion in 2014. California's most recent drought has raised concerns about the future of nut production and responsible water use in agriculture. Recent developments in genomic resources have created new breeding opportunities for water-use efficiency. Therefore, a genome wide association genetics approach was used in conjunction with new phenotyping methods introduced in the Walnut Improvement Program (WIP), at the University of California, Davis. Using 266 mature clones of 64 founders and commonly crossed cultivars, 126,654 SNPs were tested with the variation of seven ecophysiological traits, assessing water-use efficiency via carbon isotope discrimination ([delta]13C), leaf growth (%C, %N, C:N ratio), nitrogen isotope composition ([delta]15N), and leaf area traits. Gas-exchange measurements confirmed stomatal conductance as the primary mechanism of phenotypic variation. The clonal mean [delta]13C was 21.73‰, with a range of 19.32‰ to 24.41‰. A significant moderate negative correlation between %C and %N was observed (r2 = -0.38). Additionally, nitrogen related traits (%N, C:N ratio, and [delta]15N) showed highly correlated measurements between the two years, indicating a strong genetic relationship in these traits. Genotyping results showed a mean heterozygosity of 0.34 for all loci, and 0.34 heterozygosity per marker locus. Estimations of cultivar relationships were elucidated using a marker-based realized relationship matrix, confirming the accuracy of historically recorded pedigrees within the breeding program. This increased mean kinship coefficients from 0.12 to 0.78. Coefficients were used in estimations of individual trait narrow-sense heritabilities ranging from 0.27 (%C and SLA), to 0.65 (%N). Best Linear Unbiased Predictors (BLUPs) were used to as Genomic Estimated Breeding Values (GEBVs) for cultivar performance rankings to determine future crosses. Due to the complex nature of the phenotypes, and interest in decreasing bias a polygenic model was used with the realized relationship matrix, reducing false positives. Despite corrections for multiple testing, no SNPs met the threshold of significance, though prominent linkage groups differed between traits. The top 100 markers were explored for possible annotations, and 118 characterizations were confirmed across 628 unique sequences. The most abundant annotation, zinc finger protein, was associated with [delta]15N, which is consistent with other species for expression regulation and abiotic stress responses. While no specific markers were identified, this pilot study was successful at assessing the breeding potential within the WIP. Increased phenotyping in the future could lead to better prediction models and realized gain in more conscious agricultural water use in California.