Establishing a database of volatile metabolites in whole wheat (Triticum aestivum L.) flour provides novel insights into quality breeding.

Improving yield traits is the primary goal for wheat breeding, while much less attention has been paid to flavor and odor quality of wheat grain. In this study, a high-throughput protocol was developed for the identification and quantification of volatile compounds (VOCs) in wheat grains via a static head-space coupled with GC-MS. Using a diverse population consisting of 312 wheat germplasms, a database including 94 VOCs was constructed for grains. The main pathways for the biosynthesis of VOCs in wheat grains was constructed by manually assigning the volatiles to the lipoxygenase, the mevalonic acid and the shikimate/phenylalanine pathway, respectively. VOC profiling showed that grain VOCs were mainly synthesized from the LOX pathway and the content of VOCs varied largely between different wheat lines. Clear discrimination was identified between old and modern cultivars by OPLS-DA, indicating a reduction of VOCs both in species and content in modern cultivars. Correlation analysis and machine learning models established a link between VOC profiles and grain protein content. VOCs such as 1-octen-3-ol were significantly correlated with grain protein content, which can be used as diagnostic markers for wheat grain quality. In summary, our study established a comprehensive VOC database for wheat grains and explored the relationships between VOCs and grain quality traits in a diverse wheat population, providing novel insights into grain quality improvement and flavor enhancement in modern breeding programs.