|Lee, Hoonsoo - CHUNGBUK NATIONAL UNIVERSITY|
|Oh, Mirae - RURAL DEVELOPMENT ADMINISTRATION - KOREA|
|Tarpley, Lee - TEXAS AGRILIFE RESEARCH|
|Lee, Kangjin - RURAL DEVELOPMENT ADMINISTRATION - KOREA|
Submitted to: Scientific Reports
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/31/2020
Publication Date: N/A
Interpretive Summary: Measuring rice grain quality attributes is a time consuming and imprecise activity that is often necessary for breeding and genebank characterization. Visible and near infra-red (Vis/NIR) spectroscopy is a hyperspectral imaging method for collecting reflectance data on samples across a range of wavelengths that includes and extends beyond visible light into the near infra-red. Vis/NIR imaging can quickly provide a large amount of information about the physical and chemical properties of a sample. This study demonstrates that Vis/NIR imaging can classify rice grain samples according to genetic subpopulation and location where the rice was grown, classifications that cannot be made by simple visual inspection. Further, this study shows that patterns of variation from a specific range of Vis/NIR reflectance wavelengths are highly correlated with rice grain chalkiness, an important rice quality trait that influences crop value. Genome-wide analysis with over 3.2 million sequence variants revealed that genetic variation in grain chalkiness and variation in Vis/NIR reflectance are influenced by similar genomic regions indicating that both are controlled by the same genes. This suggests that Vis/NIR spectroscopy is useful as a high-throughput phenotyping tool for objective quantitative measurement of grain chalkiness. This accurate and fast measurement methodology will accelerate progress in breeding new rice varieties that produce less chalky grains. Knowledge of the specific wavelengths associated with grain chalkiness may provide insight into the biological mechanisms that are involved in this trait that strongly impacts rice grain quality.
Technical Abstract: Rice grain quality is a multifaceted quantitative trait that impacts crop value and is influenced by multiple genetic and environmental factors. Chemical, physical, and visual analyses are the standard methods for measuring grain quality. In this study, we evaluated high-throughput hyperspectral imaging for quantification of rice grain quality and classification of grain samples by genetic sub-population and production environment. Whole grain rice samples from the USDA mini-core collection grown in multiple locations were evaluated using hyperspectral imaging and compared with results from standard phenotyping. Loci associated with hyperspectral values were mapped in the mini-core with 3.2 million SNPs in a genome-wide association study (GWAS). Our results show that visible and near infra-red (Vis/NIR) spectroscopy can classify rice according to sub-population and production environment based on differences in physicochemical properties. The 702-900 nm range of the NIR spectrum was associated with the undesirable chalky grain trait. GWAS revealed that grain chalk and hyperspectral variation share genomic regions containing several plausible candidate genes for grain chalkiness. Hyperspectral quantification of grain chalk was validated using a segregating bi-parental mapping population. These results indicate that Vis/NIR can be used for non-destructive high throughput phenotyping of grain chalk and potentially other grain quality properties.