Correlation exploration of metabolic and genomic diversity in rice

Abstract

Background: It is essential to elucidate the relationship between metabolic and genomic diversity to understand the genetic regulatory networks associated with the changing metabolo-phenotype among natural variation and/or populations. Recent innovations in metabolomics technologies allow us to grasp the comprehensive features of the metabolome. Metabolite quantitative trait analysis is a key approach for the identification of genetic loci involved in metabolite variation using segregated populations. Although several attempts have been made to find correlative relationships between genetic and metabolic diversity among natural populations in various organisms, it is still unclear whether it is possible to discover such correlations between each metabolite and the polymorphisms found at each chromosomal location. To assess the correlative relationship between the metabolic and genomic diversity found in rice accessions, we compared the distance matrices for these two "omics" patterns in the rice accessions.

Results: We selected 18 accessions from the world rice collection based on their population structure. To determine the genomic diversity of the rice genome, we genotyped 128 restriction fragment length polymorphism (RFLP) markers to calculate the genetic distance among the accessions. To identify the variations in the metabolic fingerprint, a soluble extract from the seed grain of each accession was analyzed with one dimensional (1)H-nuclear magnetic resonance (NMR). We found no correlation between global metabolic diversity and the phylogenetic relationships among the rice accessions (r(s) = 0.14) by analyzing the distance matrices (calculated from the pattern of the metabolic fingerprint in the 4.29- to 0.71-ppm (1)H chemical shift) and the genetic distance on the basis of the RFLP markers. However, local correlation analysis between the distance matrices (derived from each 0.04-ppm integral region of the (1)H chemical shift) against genetic distance matrices (derived from sets of 3 adjacent markers along each chromosome), generated clear correlations (r(s) > 0.4, p < 0.001) at 34 RFLP markers.

Conclusion: This combinatorial approach will be valuable for exploring the correlative relationships between metabolic and genomic diversity. It will facilitate the elucidation of complex regulatory networks and those of evolutionary significance in plant metabolic systems.

Figure 1

Population structure of 68 rice accessions from around the world (including the 18 used in this study) based on the genotype data of 128 RFLP markers determined with Structure software.

Figure 2

Diversity of soluble metabolites among the 18 rice accessions. The global metabolic profiles of soluble metabolites were determined from 0.04-ppm integral values based on ¹H NMR spectra. The 1D ¹H NMR spectra in the range of chemical shifts from 4.29 to 0.71 ppm are shown, along with a close-up of the spectra from 3.3 ppm. Orange, japonica; blue, indica I; green, indica II.

Figure 3

Global genomic and metabolo-phenotypic diversity among the 18 rice accessions. Spearman's coefficient of correlation between the genetic and Euclidian distance matrices of the 1D NMR data was calculated. (A) Dendrogram of the phylogenetic tree obtained by the UPGMA method. (B) Metabolomic distance based on normalized data from the integral regions of the ¹H NMR spectra is shown in Additional file 3 Table S2.

Figure 4

Correlation display between genomic and metabolic diversity in rice. Top: Heatmap displaying coefficients of correlation between metabolite abundance based on the ¹H chemical shift (horizontal) and polymorphisms along each of the 12 rice chromosomes (vertical) in the 18 rice accessions. Bottom: Overlay of ¹H-NMR spectra of metabolite extracts from the 18 rice accessions. Orange, japonica; blue, indica I; green, indica II. Intra-molecular correlations against more than two NMR signals can be determined; these are allocated on the chromosomal regions in this correlation display (green and blue arrows for sucrose and lipids, respectively).