Estimation of bulk deoxynivalenol and moisture content of wheat grain samples by FT-NIR spectroscopy

When screening wheat for Fusarium head blight (FHB) resistance, cultivars are grown under heavy disease pressure and FHB disease levels are assessed using several disease evaluation criteria. The bulk deoxynivalenol (DON) levels of harvested grain samples are often used to evaluate FHB resistance of wheat varieties. In an attempt to reduce the time and cost associated with standard laboratory DON analysis, we developed calibrations for estimating bulk DON levels and moisture content (MC) of wheat grain samples using a Perkin Elmer Spectrum 400 FTIR/FTNIR spectrometer. Grain samples from nursery trials conducted to evaluate FHB resistance of several wheat varieties in 2010 and 2011 were used. Grains in a glass petri-dish were scanned in 10000 - 4000 cm ^-1 (1000-2500 nm) range. The bulk DON levels of grain samples were determined by the standard gas chromatography-mass spectrometry method while the MC was determined by oven drying method. Mean centered absorbance spectra with corresponding DON and MC values were used to develop calibration models using the partial least squares (PLS) regression technique. The DON calibration predicted the bulk DON levels with R² = 0.48, SEP = 2.4 ppm. Likewise the MC calibration predicted MC in grain samples with R² = 0.95, SEP = 0.2%. These results suggest that bulk DON level and MC of a grain sample can be non-destructively estimated in about 30 seconds per sample using FTNIR spectroscopy. This also allows adjusting the DON level of samples under a constant MC basis. This method may be used to pre-screen grain samples from scab nursery trials. High DON samples may be identified for culling while those with lower DON levels could be identified for further testing and evaluation using standard laboratory DON analysis. This will reduce the cost of FHB resistance evaluation programs.