Highly geographical specificity of metabolomic traits among Korean domestic soybeans (Glycine max)

Classification and characterization of agricultural products at molecular levels are important but often impractical with genotyping, particularly for soybeans that have numerous types of variety and landraces. Alternatively, metabolic signature, a determinant for nutritional value, can be the good molecular indicator, which reflects cultivation region-dependent factors such as climate and soil. Accordingly, we analyzed the integrative metabolic profiles of Korean soybeans cultivated in 7 different provinces (representative production areas), and explored the potential association with geographic traits. A total of 210 primary and secondary metabolites were profiled using gas-chromatography time-of-flight mass spectrometry (GC-TOF MS) and liquid-chromatography Orbitrap mass spectrometry (LC-Orbitrap MS). Despite the partial heterogeneity of the soybean varieties, the metabolomic phenotypic analysis based on multivariate statistics inferred the chemical compositional characteristics was primarily governed by the regional specificity. The OPLS-DA model proposed biomarker cluster re-composed with 5 metabolites (tryptophan, malonylgenistin, malonyldaidzin, N-acetylornithine, and allysine) (AUCs = 0.870–1.0). The most distinctive metabolic profiles were identified with the soybeans of Gunsan (middle-western coast) and Daegu (east-southern inland area), which were best characterized by the highest contents of isoflavones and amino acids, respectively. Further interrogation on geographic data suggested the combinatorial association of region-specific metabolic features with general soil texture and climate traits (total rainfall and average annual temperature).