Development of genome-wide single nucleotide polymorphism markers for variety identification of F1 hybrids in cucumber (Cucumis sativus L.)

Abstract

Plant variety identification is essential for the official registration of new cultivars and for variety protection. It is therefore imperative to have rapid and highly reliable methods for plant variety identification. In this study, a core 96-SNP set optimized for use with Fluidigm EP1TM assay was developed to facilitate cultivar identification in cucumber. Genotyping-by-sequencing (GBS) of 88 F1 hybrids comprising nine cultivar groups (Korean Gasi, Korean Nakhap, Korean Baekdadagi, Japanese Gasi, Chinese Gasi, Chinese Baekdadagi, Southeast Asian slice, European slice, and pickle groups) was carried out, and a total of 10,996 high-quality SNPs were generated. Population structure, principle component analysis, and hierarchical clustering indicated that the nine cultivar groups were clearly divided into four subpopulations (subpopulation G for Gasi type, B for Baekdadagi type, N for Nakhap type, and E for slice and pickle types). Population differentiation measures based on pairwise FST values indicated that the closest relationship (0.124) was between subpopulations B and N and subpopulations E (0.140) and G, while subpopulations B and E were the most clearly distinct from each other (0.208). The metrics of genetic differentiation measured by AMOVA also indicated that significant genetic division exists among the subpopulations, while variation among cultivars within each subpopulation is extremely low, indicating high genetic exchange within the subpopulations. By applying filtering criteria to the SNPs, including FST and linkage disequilibrium decay values, 240 SNPs were filtered from the original 10,996 for use with a Fluidigm assay, from which a core 96-SNP set was obtained. This set demonstrated high reproducibility and discrimination power. Hierarchical clustering confirmed that the core 96-SNP set could distinguish all 88 cultivars and delineate them clearly according to their population structures. A Mantel test based on 10,996 SNPs and 37 traits revealed a high level of correlation (r = 0.57) between molecular and morphological distance. Our core 96-SNP set is not only suitable for generating signature profiles of released cucumber cultivars, but also represents a promising supplement to morphological trait-based distinctness, uniformity, and stability testing for cultivar identification.