Genome-Wide Association Study of Seed Folate Content in Common Bean

Abstract

Plant-derived folates (Vitamin B₉) are essential components of the human diet. They provide one-carbon units that are required for the synthesis of nucleic acids and proteins, and folate deficiency is associated with numerous adverse health conditions. The development of high-folate cultivars of common bean (Phaseolus vulgaris L.) and other staple crops is an important tool to combat folate deficiency. A population of 96 P. vulgaris accessions, representing major North American market classes, was grown in 2 years in Ontario, Canada. The population was genotyped for 5,361 molecular markers with an Illumina Infinium platform. Total folate was extracted from mature seeds using the tri-enzyme extraction method and quantified based on a microbiological assay with Lactobacillus rhamnosus. Significant genetic diversity for folate content was observed among the population in both years of study, and folate content had a range 113-222 μg per 100 g of seeds. Quantitative trait loci (QTL) for seed folate content were identified based on a genome-wide association study (GWAS). Six QTL were identified on Chr. 4, 6, 8, and 11, with three in each year of field trials. Both QTL on Chr. 11 occurred in genomic regions that were syntenic to seed folate QTL detected in previous work with P. vulgaris, Z. mays, and O. sativa. Candidate genes were identified for these QTL that might be targets for the development of molecular markers for selecting P. vulgaris cultivars with improved seed folate content. This work reports the largest survey of genetic diversity for seed folate content in P. vulgaris and identified several genotypes, including SCN4, Bat 93, OAC Redstar, and Pompadour 1014, that would be useful for breeding beans with higher than average folate levels.

Keywords: GWAS; QTL; folate; nutrition; vitamin B9.

FIGURE 1

Frequency histograms of total seed folate content in the diversity panel evaluated in 2015 and 2016. The counts for total seed folate are based on the least squared means from the separate ANOVAs for the experiments conducted in 2015 (A) and 2016 (B).

FIGURE 2

Mean total seed folate of P. vulgaris accessions among market classes and subpopulations. Least squared means of market classes are presented in panel (A) and subpopulations, defined by the K Groups, are presented in panel (B). Error bars represent SEM of two biological replications of P. vulgaris accessions grown in a RCBD across 2 years. Significant differences are represented by letter codes above bars, and Tukey’s adjustment was applied to the PROC MIXED ANOVA in SAS (Cary, NC).

FIGURE 3

Hierarchical clustering of P. vulgaris accessions in the diversity panel based on 2,522 SNP markers. The dendrogram was generated in Flapjack using the default parameters (Milne et al., 2010).

FIGURE 4

Principal component analysis of 96 P. vulgaris accessions based on 2,522 SNP markers. In panel (A), the Mesoamerican accessions are colored blue and the Andean accessions are colored green. In panel (B), the accessions are colored based on the K groups from Supplementary Figure 2 and Figure 3. The PCA plot was visualized with CurlyWhirly (Milne et al., 2014). PC1, the first principal component; PC2, the second principal component; PC3, the third principal component.

FIGURE 5

Manhattan plots of QTL for total seed folate content in the diversity panel that were detected by GWAS with the FarmCPU model in rMVP. (A) Plots based on 2015 data. (B) Plots based on 2016 data. The significance threshold is indicated by a dashed line on each plot, and markers above the significance threshold are colored red. SNP density is indicated in the legends to the right of each plot.

FIGURE 6

Box plots showing the phenotypic distributions of total seed folate content for six QTL identified by GWAS with the FarmCPU model. The LSMEANS from the separate ANOVAs for the 2015 and 2016 experiments were plotted. Accessions that were heterozygous for the QTL marker were not included in the analysis. Each boxplot is labeled with the SNP genotype and the year (e.g., GG2015). The chromosomes and physical positions of the SNPs are indicated below each plot.

FIGURE 7

Box plots showing the phenotypic distributions of total seed folate contents for haplotypes based on six QTL identified by GWAS with the FarmCPU model. The QTL names are indicated above each panel (A–H). The haplotypes containing the QTL allele with the highest average seed folate content are underlined, and the SNP alleles are identified by black arrowheads. Each haplotype is followed by either 15 or 16, which corresponds to the 2015 and 2016 experiments, respectively.