Genome-Wide Association Analysis of Mucilage and Hull Content in Flax (Linum usitatissimum L.) Seeds

Abstract

New flaxseed cultivars differing in seed mucilage content (MC) with low hull content (HC) represent an attractive option to simultaneously target the food and feed markets. Here, a genome-wide association study (GWAS) was conducted for MC and HC in 200 diverse flaxseed accessions genotyped with 1.7 million single nucleotide polymorphism (SNP) markers. The data obtained for MC and HC indicated a broad phenotypic variation and high (~70%) and a moderate (~49%) narrow sense heritability, respectively. MC and HC did not differ statistically between fiber and oil morphotypes, but yellow-seeded accessions had 2.7% less HC than brown-seeded ones. The genome-wide linkage disequilibrium (LD) decayed to r2 = 0.1 at a physical distance of ~100 kb. Seven and four quantitative trait loci (QTL) were identified for MC and HC, respectively. Promising candidate genes identified include Linum usitatissimum orthologs of the Arabidopsis thaliana genes TRANSPARENT TESTA 8, SUBTILISIN-LIKE SERINE PROTEASE, GALACTUROSYL TRANSFERASE-LIKE 5, MUCILAGE-MODIFIED 4, AGAMOUS-LIKE MADS-BOX PROTEIN AGL62, GLYCOSYL HYDROLASE FAMILY 17, and UDP-GLUCOSE FLAVONOL 3-O-GLUCOSYLTRANSFERASE. These genes have been shown to play a role in mucilage synthesis and release, seed coat development and anthocyanin biosynthesis in A. thaliana. The favorable alleles will be useful in flaxseed breeding towards the goal of achieving the ideal MC and HC composition for food and feed by genomic-based breeding.

Resequencing 200 Flax Cultivated Accessions Identifies Candidate Genes Related to Seed Size and Weight and Reveals Signatures of Artificial Selection

“…The population genetic structure is basically consistent with the morphological differences of cultivars. △K shows that the most likely group number is K = 2, which is consistent with previous studies (Soto-Cerda et al, 2018). Oil flax is basically clustered into one group and has a long genetic distance from OF and fiber groups ( Figures 1B, E).…”

“…Association mappings based LD have provided high resolution for exploring genes that may sustain variability in phenotypes. The extent of LD reflects the requirement of marker density and mapping resolution in GWAS (Soto-Cerda et al, 2018). The LD extent between molecular markers is mostly influenced by recombination rate, genetic isolation, population sub-structure, selection effect, and other random factors, and the mating system of species also has a decisive influence on the LD extent (Rafalski and Morgante, 2004).…”

“…In addition, a total of 17,288 SNPs were employed to perform GWAS for 11 traits in three different biparental flax mapping populations composed of 260 lines, and 33 quantitative trait loci (QTL) were detected (You et al, 2018a). Soto-Cerda et al (2018) used 771,914 filtered SNPs to perform GWAS for mucilage (MC) and hull content (HC) in flax seeds, and seven and four QTLs were identified for MC and HC, respectively. Nevertheless, these studies were based on SSR and Reduced-Representation Genome Sequencing such as SLAF and GBS, which are far less efficient and accurate than re-sequencing.…”

Mapping Quantitative Trait Loci onto Chromosome-scale Pseudomolecules in Flax

“…The coordinates of SNPs based on the PCPs in two publications [27,28] were first retrieved for their scaffold names and corresponding coordinates on the scaffolds because these SNPs were initially identified from scaffold sequences. These SNPs were then converted to the RCPs using Program S3.…”

“…These QTL were identified mainly using SSR or SNP markers ( Table 2). Most (200) of the QTL were mapped based on genetic maps [15,[18][19][20][21][22][23][24], scaffold sequences [17,25,26], or an early (hereafter pre-released) version of chromosome-scale pseudomolecules (PCPs) [27,28]; however, only 67 QTL for pasmo severity (PAS) were mapped on the most recent release of the chromosome-scale pseudomolecules (RCPs) ( Table 2) [14,29]. For comparison purposes, coordinates of the QTL based on the PCPs must be converted to the RCPs [29] because the two versions are slightly different.…”