Genome-Wide Single Nucleotide Polymorphism Discovery and the Construction of a High-Density Genetic Map for Melon (Cucumis melo L.) Using Genotyping-by-Sequencing.

Che-Wei Chang,Chih-Wei Tung,Yu-Hua Wang

doi:10.3389/fpls.2017.00125

Abstract

Although genotyping-by-sequencing (GBS) enables the efficient and low-cost generation of large numbers of markers, the utility of resultant genotypes are limited, because they are enormously error-prone and contain high proportions of missing data. In this study, we generated single nucleotide polymorphism (SNP) markers for 109 recombinant inbred lines of melon (Cucumis melo L.) using the GBS approach and ordered them according to their physical position on the draft double haploid line DHL92 genome. Next, by investigating associations between these SNPs, we discovered that some segments on the physical map conflict with linkage relationships. Therefore, to filter out error-prone loci, 4,110 SNPs in which we have a high degree of confidence were selected as anchors to test independence with respect to unselected markers, and the resultant dataset was then analyzed using the Full-Sib Family Haplotype (FSFHap) algorithm in the software TASSEL 5.2. On the basis of this analysis, 22,933 loci that have an average rate of missing data of 0.281% were used to construct a genetic map, which spans 1,088.3 cM across 12 chromosomes and has a maximum spacing of 6.0 cM. Use of this high-quality linkage map enabled the identification of several quantitative trait loci (QTL) known to control traits in fruit and validated our approach. This study highlights the utility of GBS markers for the identification of trait-associated QTLs in melon and facilitates further investigation of genome structure.

Highlights

Melon (Cucumis melo L., 2n = 24) is an economically very important fruit crop
46,590 single nucleotide polymorphism (SNP) sites were identified in biparental RILs, the volume of missing data could significantly hinder application of this dataset, and the alignment of sequenced reads to the DHL92 draft genome (Garcia-Mas et al, 2012; Argyris et al, 2015) may affect the accuracy of SNP position
To improve marker quality and accuracy for quantitative trait loci (QTL) mapping, we initially investigated the linkage relationships of 4,164 selected SNPs that did not have any missing data in RILs as well as the two parents and calculated pairwise logarithm of odds (LOD) scores and recombination fraction. While these 4,164 markers were classified into 12 linkage groups with LOD scores greater than 10, 54 could not be clustered correctly (Figure 1 and Supplementary Table S1); these misallocated markers were removed and the remaining 4,110 used to construct a reference-based and a de novo map by ordering markers on the basis of the physical position of SNPs and the minimum spanning tree (MST) method (Wu et al, 2008), respectively

Summary

Introduction

The nutritional quality of melon is high, as these fruits are low in calories and rich in nutrients including potassium, vitamin C, and β-carotene (Lester, 1997). While it is debated whether melons first evolved in Africa (Kerje and Grum, 2000) or in Asia (Sebastian et al, 2010), this fruit is highly diverse in morphology and biochemistry (Stepansky et al, 1999) and is classified into two subspecies, melo and agrestis (Jeffrey, 1980). Connecting physical and high-density genetic maps enables more precise positional detection of recombination breakpoints (Yu et al, 2011) and the identification of hot and cold recombination spots (Spindel et al, 2013)

Methods

Results

Conclusion