Non destructive estimation of total phenol and crude fiber content in intact seeds of rapeseed–mustard using FTNIR

Abstract

In the present study calibration models were developed for non destructive estimation of total phenol and crude fiber content in intact rapeseed mustard seed by Fourier transform near infrared spectroscopy (FTNIR). Rapeseed–mustard (Brassica spp.) is an important group of oilseed crops in India. The defatted meal that is left after oil extraction is a high value by-product for animal nutrition and is also a potential source of protein for human nutrition. However the utilization of mustard meal in animal (monogastric livestock such as poultry and pigs) and human nutrition is limited due to high content of antinutritive compounds such as fiber and phenolic. In order to develop genotypes with low or high contents of phenol and fiber fast screening of existing genotypes is required. The traditional methods for estimation of total phenol and crude fiber are destructive and time consuming. A total of 115 rapeseed mustard genotypes were quantitatively analyzed for total phenol and crude fiber content by wet chemical methods, their Fourier transform near infrared spectra was correlated to resulting data by means of partial least square regression and calibration models were developed. The optimal models were achieved with coefficient of determination (R) 2 of 0.96 and 0.91 and root mean square error of cross validation (RMSECV) of 0.08 and 0.41 with residual predictive deviation (RPD) values of 4.98 and 3.37 for total phenol and crude fiber, respectively. Test validation resulted in RMSEP of 0.11 for total phenol and RMSEP of 0.28 was observed for crude fiber content.