Genome-Wide Association Study of Waterlogging Tolerance in Barley (Hordeum vulgare L.) Under Controlled Field Conditions

Abstract

Waterlogging is one of the main abiotic stresses severely reducing barley grain yield. Barley breeding programs focusing on waterlogging tolerance require an understanding of genetic loci and alleles in the current germplasm. In this study, 247 worldwide spring barley genotypes grown under controlled field conditions were genotyped with 35,926 SNPs with minor allele frequency (MAF) > 0.05. Significant phenotypic variation in each trait, including biomass, spikes per plant, grains per plant, kernel weight per plant, plant height and chlorophyll content, was observed. A genome-wide association study (GWAS) based on linkage disequilibrium (LD) for waterlogging tolerance was conducted. Population structure analysis divided the population into three subgroups. A mixed linkage model using both population structure and kinship matrix (Q+K) was performed. We identified 17 genomic regions containing 51 significant waterlogging-tolerance-associated markers for waterlogging tolerance response, accounting for 5.8-11.5% of the phenotypic variation, with a majority of them localized on chromosomes 1H, 2H, 4H, and 5H. Six novel QTL were identified and eight potential candidate genes mediating responses to abiotic stresses were located at QTL associated with waterlogging tolerance. To our awareness, this is the first GWAS for waterlogging tolerance in a worldwide barley collection under controlled field conditions. The marker-trait associations could be used in the marker-assisted selection of waterlogging tolerance and will facilitate barley breeding.

Keywords: barley; candidate genes; genome-wide associated study; marker-trait association; quantitative trait loci; waterlogging tolerance.

Figure 1

Model-based populations of spring barley collection: (A) Two-dimension distribution analyzed by principal component analysis (PCA) by JMP Genomics 9.1, (B) phylogenetic tree constructed by neighbor-joining (NJ) of genetic distance by TASSEL 5.2.28, and (C) Classification of three populations using STRUCTURE 2.3.4. The color code indicates the distribution of the accessions to different populations (Q1: red, Q2: green, Q3: blue) consistent in (A–C).

Figure 2

Plot of pairwise SNP linkage disequilibrium (LD) r² value as a function of inter-marker genetic distances (Mbp) of 247 spring barley genotypes. The blue curve represents the smoothing spline regression model fit to LD decay. The red line represents the baseline r² value at 0.1. The intersection of the fitted smoothing spline and r² was observed at around 1,460,356 bp.

Figure 3

Manhattan plots resulting from the SNP-based GWAS in overall control under field conditions. Manhattan plots for Biomass (BIO), Spikes per plant (SP), Grains per plant (GP), Kernel weight per plant (KWP), and Plant height (PH) are shown in (A–E), respectively, and the x-axis shows SNP loci along the seven barley chromosomes. The horizontal red line shows the genome-wide significance threshold P-value of 1.6 × 10⁻⁴ or –log₁₀ (P-value) value of 3.8. GWAS was performed using the MLM (Q + K) model in JMP Genomics for the field traits.

Figure 4

Manhattan plots resulting from the SNP-based GWAS in waterlogging treatment under field conditions. Manhattan plots for Biomass (BIO), Spikes per plant (SP), Grains per plant (GP), Kernel weight per plant (GWP), Plant height (PH), and Waterlogging score (WLS) are shown in (A–F), respectively, and the x-axis shows SNP loci along the seven barley chromosomes. The horizontal red line shows the genome-wide significance threshold P-value of 1.6 × 10⁻⁴ or –log₁₀ (P-value) value of 3.8. GWAS was performed using the MLM (Q + K) model in JMP Genomics for the field traits.

Figure 5

Manhattan plots resulting from the SNP-based GWAS identified in the relative dataset. Manhattan plots for Biomass (BIO), Spikes per plant (SP), Grains per plant (GP), Kernel weight per plant (GWP), Plant height (PH), and Waterlogging score (WLS) are shown in (A–F), respectively, and the x-axis shows SNP loci along the seven barley chromosomes. The horizontal red line shows the genome-wide significance threshold P-value of 1.6 × 10⁻⁴ or –log₁₀ (P-value) value of 3.8. GWAS was performed using the MLM (Q + K) model in JMP Genomics for the field traits.