Abstract
Gluten composition is an important quality parameter for wheat flour, because it is strongly correlated to baking quality. Wheat proteins are commonly extracted stepwise and analysed using RP-HPLC-UV to determine the gluten composition. This procedure is very time-consuming and labour-intensive. Therefore, a new, fast and easy method to quantitate gluten proteins was established using NIR spectroscopy (NIRS). PLS-regression models were calculated containing 207 samples for calibration and 169 for test set validation. Albumin/globulin (ALGL), gluten, gliadin and glutenin content was predicted with a root mean square error of prediction (RMSEP) of 2.01 mg/g, 6.09 mg/g, 4.25 mg/g and 3.50 mg/g, respectively. High-molecular-weight glutenin subunits (HMW-GS) and low-molecular-weight glutenin subunits (LMW-GS) were predicted with a RMSEP of 1.12 mg/g and 2.38 mg/g. The relative error was too high for ALGL, LMW-GS and HMW-GS, but that of gluten, gliadins and glutenins was in a range comparable to the reference method. Therefore, the new NIRS method can be used to estimate the gluten composition of wheat flour, including the gliadin/glutenin and the LMW-GS/HMW-GS ratio.