Near-Infrared Reflectance Spectroscopic Analysis of Moisture, Fat, Protein, and Physiological Activity in Buckwheat Flour for Breeding Selection

We examined the feasibility of near-infrared (NIR) reflectance spectroscopy to estimate the contents of moisture, fat, and protein in the flours of common and tartary buckwheat harvested in 1996 and 1997. Multiple linear -regression analysis for NIR spectral and chemical data was carried out. The standard error of the prediction (SEP) for the samples harvested in 1996 was 0.22% for moisture, 0.06% for fat, and 0.21% for protein, when the calibration equations obtained from the analysis of common and tartary buckwheat were used. When the same calibration equations for the buckwheat flour harvested in 1997 were used, the correlation graphs between the reference values and the calculated values were skewed and biased. However, the relative values were sufficient for the selection based on the contents of the constituents, and the bias-corrected SEPs were 0.15% for moisture, 0.10% for fat, and 0.44% for protein. We also examined the feasibility of NIR spectroscopy to estimate the total polyphenol content and 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical-scavenging activity in the flours of common buckwheat harvested in 1998. The estimation error was acceptable for use in the selection of varieties with a higher activity. Using the NIR method, the contents of moisture, fat and protein, and physiological activity (total polyphenol content and DPPH radical-scavenging activity) in the buckwheat flours could be successfully estimated for a simple and rapid-breeding selection