Estimating the fatty acid composition of the oil in intact-seed rapeseed (Brassica napus L.) by near-infrared reflectance spectroscopy

The objective of this work was to evaluate the potential of near-infrared reflectance spectroscopy (NIRS) as a rapid method to estimate the fatty acid composition of the oil in intact-seed samples of rapeseed. A total of 549 samples (3 g intact seed) from selected mutant and breeding lines were scanned by NIRS, and 220 of them were selected and scanned again by using two different adapters, which reduced the sample size to 300 and 60 mg, respectively. Selected samples were analysed by gas liquid chromatography and calibration equations for individual fatty acids were developed. Calibrations for oleic, linoleic, linolenic, and erucic acid were highly accurate, with values of r(2) in cross validation from 0.95 to 0.98 (samples of 3 g), from 0.93 to 0.97 (300 mg), and from 0.84 to 0.96 (60 mg). Calibrations for palmitic and stearic acid were less accurate, with values of r(2) in cross validation always lower than 0.8, probably because of the narrow range available for these fatty acids. The accuracy of the calibration equations for eicosenoic acid was very low (r(2) = 0.69 in 3 g samples), although improved equations were developed (r(2) from 0.78 to 0.91) when the relationship between erucic and eicosenoic acid was taken into account. We conclude that NIRS is a powerful technique to estimate the fatty acid composition of the oil in rapeseed, provided that samples covering a wide range of fatty acid levels are available, with the advantage that such estimation is possible with few additional costs when NIRS is used for the determination of other seed quality traits.