Assessment of sesame oil fatty acid and sterol composition with FT-NIR spectroscopy and chemometrics

In this study the Fourier transform-near infrared (FT-NIR) spectroscopy technique was used to predict the fatty acid and sterol content of sesame. For this purpose, partial least square regression (PLS-R)-based prediction models were developed, which relate the FT-NIR spectra to reference GC measurements. In total, 39 sesame samples were collected from local producers around Mugla Province. Sesame oil was extracted from the seeds by using a screw press and extracted oil samples were analyzed without any refining. The results showed that among different fatty acids found in sesame oil, oleic and linoleic acid contents (which account for 85% of the total fatty acids) can be precisely predicted with corresponding PLS-R models having R 2 = 0.991, RMSECV = 0.092%, RPD = 10.7 and R 2 = 0.988, RMSECV = 0.118, RPD = 9.01, respectively. Similarly, good model performances were obtained for the fatty acids grouped according to their saturation degree, namely saturated, monounsaturated, and polyunsaturated fatty acids with R 2 , RMSECV, and RPD values in the ranges of 0.865-0.976, 0.148%-0.148%, and 2.72-6.41, respectively. In addition, models with moderate quality for the s-sitosterol and Δ5-avenasterol content of the sesame oils could be established with R 2 = 0.756, RMSECV = 0.651%, RPD = 2.04 for s-sitosterol and R 2 = 0.823, RMSECV = 0.343%, RPD = 2.38 for Δ5-avenasterol content models. In conclusion, FT-NIR spectroscopy was proved to be a valuable analytical technique that enables rapid and simultaneous measurement of the major lipid constituents of sesame oil, which can be used effectively in quality control and breeding studies of sesame.