Genome-Wide Association Study Reveals Novel Marker-Trait Associations (MTAs) Governing the Localization of Fe and Zn in the Rice Grain

Abstract

Micronutrient malnutrition due to Fe and Zn, affects around two billion people globally particularly in the developing countries. More than 90% of the Asian population is dependent on rice-based diets, which is low in these micronutrients. In the present study, a set of 192 Indian rice germplasm accessions, grown at two locations, were evaluated for Fe and Zn in brown rice (BR) and milled rice (MR). A significant variation was observed in the rice germplasm for these micronutrients. The grain Fe concentration was in the range of 6.2-23.1 ppm in BR and 0.8-12.3 ppm in MR, while grain Zn concentration was found to be in the range of 11.0-47.0 ppm and 8.2-40.8 ppm in the BR and MR, respectively. Grain Fe exhibited maximum loss upon milling with a mean retention of 24.9% in MR, while Zn showed a greater mean retention of 74.2% in MR. A genome-wide association study (GWAS) was carried out implementing the FarmCPU model to control the population structure and kinship, and resulted in the identification of 29 marker-trait associations (MTAs) with significant associations for traits viz. FeBR (6 MTAs), FeMR (7 MTAs), ZnBR (11 MTAs), and ZnMR (5 MTAs), which could explain the phenotypic variance from 2.1 to as high as 53.3%. The MTAs governing the correlated traits showed co-localization, signifying the possibility of their simultaneous improvement. The robust MTAs identified in the study could be valuable resource for enhancing Fe and Zn concentration in the rice grain and addressing the problem of Fe and Zn malnutrition among rice consumers.

Keywords: Fe; GWAS; Zn; biofortification; brown rice; donors; milled rice; rice.

FIGURE 1

Population structure of the association mapping panel comprising of 192 genotypes analyzed by model based and PCA based approaches. (A) ΔK plot depicting three subgroups in the population by Evanno’s method. The highest ΔK was 580 at K = 3, (B) the bar plot showing the three sub-populations identified. POP2 was the largest and showed remarkable admixture with POP1. POP3 was the smallest group, which showed less admixing with POP1 (C) 3D graph depicting the distribution of accessions along the three PCs (D) scree plot depicting the number of significant PCs. There were three PCs that explained a cumulative variation of ∼40%.

FIGURE 2

Linkage disequilibrium (LD) decay plot of the association mapping panel derived from 31,132 SNPs, plotted against physical distance in base pairs (bp) and co-efficient of determination (r²).

FIGURE 3

Boxplots showing distribution of grain Fe and Zn content among the sub-populations POP1, POP2, and POP3. FeBR, Fe concentration in brown rice; FeMR, Fe concentration in milled rice; ZnBR, Zn concentration in brown rice; ZnMR, Zn concentration in milled rice.

FIGURE 4

Correlations of Fe and Zn between the locations (diagonal elements) and independent locations at New Delhi (upper diagonal) and Aduthurai, Tamil Nadu (lower diagonal). FeBR, Fe concentration in brown rice; FeMR, Fe concentration in milled rice; ZnBR, Zn concentration in brown rice; ZnMR, Zn concentration in milled rice.

FIGURE 5

Manhattan plots and Q–Q plots derived through FarmCPU model depicting the significant marker trait associations on 12 chromosomes of rice for the traits analyzed at New Delhi (A) and Aduthurai (B) locations.