2019 ASHS Annual Conference

Quality in horticultural crops is often the result of many subjective characteristics working together harmoniously. Improving quality through marker-assisted selection would therefore seem unlikely to work. Here we report an approach that can do so. While Next-Gen sequencing is affordable, these complex trait analyses are limited by the challenge of integrating and analyzing large dataset of descriptive information (phenomics): much software is designed to process only small numbers of molecular markers, analytic pipelines can only analyze a few traits, and different data sets are difficult to combine. Several techniques address these challenges. We evaluated 36 horticultural-quality traits in a large (N=175) mapping population generated from a cross of Chinese kale and broccoli. Phenomic data were captured with tools developed from the One Handheld Per Breeder Initiative and aggregated in a database. We developed a pipeline to generate high quality and density genome wide markers using genotype-by-sequencing across new accurate reference genomes. To interpret these high-dimensional datasets, we modified R/qtl2 mapping software to determine QTL confidence intervals and assign likelihoods to 290 putative genes for quality-related phenotypes. Combinations of these traits were associated with overall quality, and sometimes linked genetically. These can be used for marker-assisted selection. A similar approach can be used to improve the composition of secondary metabolites, morphology, and environmental stress response.