Genotyping-by-sequencing of Brassica oleracea vegetables reveals unique phylogenetic patterns, population structure and domestication footprints.

Brassica oleracea forms a diverse and economically significant crop group. Improvement efforts are often hindered by limited knowledge of diversity contained within available germplasm. Here, we employ genotyping-by-sequencing to investigate a diverse panel of 85 landrace and improved B. oleracea broccoli, cauliflower, and Chinese kale entries. Ultimately, 21,680 high-quality SNPs were used to reveal a complex and admixed population structure and clarify phylogenetic relationships among B. oleracea groups. Each broccoli landrace contained, on average, 8.4 times as many unique alleles as an improved broccoli and landraces collectively represented 81% of all broccoli-specific alleles. Commercial broccoli hybrids were largely represented by a single subpopulation identified within a complex population structure. Greater allelic diversity in landrace broccoli and 96.1% of SNPs differentiating improved cauliflower from landrace cauliflower were common to the larger pool of broccoli germplasm, supporting a parallel or later development of cauliflower due to introgression events from broccoli. Chinese kale was readily distinguished by principal coordinate analysis. Genotyping was accomplished with and without reliance upon a reference genome producing 141,317 and 20,815 filtered SNPs, respectively, supporting robust SNP discovery methods in neglected or unimproved crop groups that lack a reference genome. This work clarifies the population structure, phylogeny, and domestication footprints of landrace and improved B. oleracea broccoli using many genotyping-by-sequencing markers. Additionally, a large pool of genetic diversity contained in broccoli landraces is described which may enhance future breeding efforts.