Introduction of fast, cheap and reliable methods in the analysis of agricultural products is of great importance for the breeding programs targeting to enhance quality traits. The objective of this study was to develop a NIRs (Near Infrared Reflectance Spectroscopy) calibration to enable a fast evaluation of quality traits in wheat flour. We used 120 flour samples from 40 bread wheat varieties as plant material. We developed calibrations based on Partial Least Squares (PLS) regression model for moisture, protein ratio, gluten ratio, gluten index value, sedimentation value, and modified sedimentation value. In the generated calibrations, it was found that the correlation coefficients between the reference analyses and NIR readings were higher than 0.90 for all variables (r=0.92 for moisture ratio, r=0.90 for protein ratio, r=0.94 for gluten ratio, r=0.93 for sedimentation value and r=0.92 for modified sedimentation value) except gluten index value(r=0.87). It was determined that NIR estimates of moisture, protein and gluten were in acceptable levels when the standard error of estimate (SEE), root mean square error of prediction (RMSEP) and Root-Mean-Square Error of Cross-Validation (RMSECV) in the developed calibrations were taken into account. Relative deviation prediction (RPD) values indicated that calibration model for gluten ratio had the potential to be used for screening (RPD=3.25). The other parameters (RPD values between 2 and 3) except gluten index value (RPD=1.90) had rough screening potential. Regression coefficients (R 2 ) of calibration parameters were above 0.80, except for gluten index value. Furthermore, as a result of the correlation analysis performed to determine the relationships of the scanned interval (1200-2400 nm) with the investigated traits, it was found that the variables except gluten index value and moisture level had high correlation values in the similar absorption regions. Ten wavelength regions had similar direction of correlation coefficient between the 1200-2400 nm for these traits. General results of this study suggest that the calibration models developed here could successfully estimate most of the wheat flour quality traits, whereas better models are needed for more robust estimations on some other traits, such as gluten index value.

Protein, Gluten, Triticum aestivum

Presentation: oral