Genome-wide association mapping of Hagberg falling number, protein content, test weight, and grain yield in U.K. wheat

Abstract

Association mapping using crop cultivars allows identification of genetic loci of direct relevance to breeding. Here, 150 U.K. wheat (Triticum aestivum L.) cultivars genotyped with 23,288 single nucleotide polymorphisms (SNPs) were used for genome-wide association studies (GWAS) using historical phenotypic data for grain protein content, Hagberg falling number (HFN), test weight, and grain yield. Power calculations indicated experimental design would enable detection of quantitative trait loci (QTL) explaining ≥20% of the variation (PVE) at a relatively high power of >80%, falling to 40% for detection of a SNP with an R² ≥ .5 with the same QTL. Genome-wide association studies identified marker-trait associations for all four traits. For HFN (h ² = .89), six QTL were identified, including a major locus on chromosome 7B explaining 49% PVE and reducing HFN by 44 s. For protein content (h ² = 0.86), 10 QTL were found on chromosomes 1A, 2A, 2B, 3A, 3B, and 6B, together explaining 48.9% PVE. For test weight, five QTL were identified (one on 1B and four on 3B; 26.3% PVE). Finally, 14 loci were identified for grain yield (h ² = 0.95) on eight chromosomes (1A, 2A, 2B, 2D, 3A, 5B, 6A, 6B; 68.1% PVE), of which five were located within 16 Mbp of genetic regions previously identified as under breeder selection in European wheat. Our study demonstrates the utility of exploiting historical crop datasets, identifying genomic targets for independent validation, and ultimately for wheat genetic improvement.

FIGURE 1

Pearson correlation between grain yield (Yield), the presence or absence of the wheat/rye chromosome 1B/1R translocation, grain protein content (PRT), Hagberg falling number (HFN), and test weight (TW)

FIGURE 2

Population subsubstructure in the association mapping panel of 150 accessions, based on (a) all markers (n = 16,801), and (b) skimmed markers (n = 463; linkage disequilibrium threshold R²  = .2). Markers with a minor allele frequency >0.05 were used for both analyses. The presence or absence of the chromosome 1B/1R wheat/rye chromosomal substitution is indicated by red circles and blue triangles, respectively. PC, principal component

FIGURE 3

Manhattan plots of genome‐wide association studies for Hagberg falling number (HFN; a), protein content (PRT; b), test weight (TW; c), and grain yield (Yield; d). Markers are shown in genetic map order, according to the genetic map published by Wang et al. (2014). Unmapped markers are not shown here but are listed in Supplemental Table S5. The significance threshold (−log₁₀ P = 3) is indicated by the horizontal red line

FIGURE 4

Histogram of Hagberg falling number (HFN; a), protein content (PRT; b), test weight (TW; c), grain yield (YLD; d) in the association mapping panel