The development of near infrared (NIR) spectroscopy calibrations for the prediction of wheat and flour quality

Abstract

The protein component of wheat and wheat flour is widely recognised as having unique properties that result in the characteristic textures of the wide range of baked products produced throughout the world. A wide range of techniques have been applied to the study of these characteristics and recently a number of approaches using the size-exclusion high performance liquid chromatography (SE-HPLC) technique have been reported. One method in particular, Profilblé®, developed in France, has been applied to the analysis of UK and French-grown wheats in work reported here.

Near infrared (NIR) spectroscopy is widely used in the global cereals industry for the assessment of compositional characteristics such as protein, moisture and lipid contents. As the use of more sophisticated instrumentation and data processing approaches has become more widespread, however, so too has the development of new, more advanced applications. This study was undertaken to develop NIR calibrations for a number of measures of wheat and flour quality in general and protein quality in particular.

New NIR calibrations have been developed using a research grade scanning instrument for over 300 samples of wheat from three harvest years (1999-2001) grown in both the UK and France. Calibrations for compositional parameters gave very high levels of performance as would be expected. Calibrations for both test baking parameters and measures of protein quality, particularly those derived from SE-HPLC analysis, also indicated potential for routine use. When the calibrations were assessed using a separate validation set, the performance of those for protein quality attributes were consistent with the results for the calibration set, while those for test baking parameters were poor, indicating limited potential for routine use. Calibrations developed using a dataset combining samples analysed in both the UK and France exhibited similar levels of accuracy to those from UK-analysed samples alone, although they promised to be more robust in routine use.

Overall, UK-grown samples tended to have a balance of gluten properties in which the proportion of gliadin material was greater than those from France. The ranking of performance for varieties analysed in both locations using different test baking regimes was very similar, however, the differences in gluten properties observed notwithstanding. There was some evidence, however, that the performance of flour in the French breadmaking test was not so reliant on higher levels of protein content as was generally the case for Chorleywood Bread Process assessment in the UK.

The new calibrations are now commercially available for users and further information about them may be obtained from the author (s.millar@campden.co.uk).