Influência da temperatura na composição dos ácidos graxos e caracterização da qualidade e do teor de óleo de genótipos de girassol por espectroscopia de NIR

Abstract

This work aimed evaluate the influence of temperature on fatty acid composition in sunflower oil grown in low latitudes, specifically to different environmental conditions of Brazil (0 °S to 23 °S); differentiating sunflower genotype with different oleic and linoleic fatty acid contents by discriminant analysis using Near-infrared reflectance (NIR) spectroscopy and establish a calibration model using data from the near infrared reflectance spectroscopy (NIR) and evaluate the oil content of sunflower seeds produced under various environmental conditions in Brazil. In study of environmental influences the composition of fatty acids was observed that the high oleic genotypes showed good stability in high levels of oleic and linoleic acids, regardless of location. In conventional genotypes, there was a large variation in the average contents of these acids, depending on the minimum temperature and there was a more important effect of the environment, compared to the effect of genotype. For conventional genotypes, the increase in oleic acid reflected in a decrease in palmitic and linoleic acids. However, no associations were observed between stearic acid and other fatty acids. The use the NIR spectroscopy for differing genotypes with oleic and linoleic fatty acid contents, were established curves of calibration using principal component analysis. The clustering precision was determined by d, the Mahalanobis generalized distance between the genotype and the closest group, and D, the distance up to the next group. The genotypes were well classified when d was close to 1 and the ratio D/d was close to or higher than 2. The discriminant analysis allowed the classification of the high oleic and mid-oleic genotypes even though they had similar signals in gas chromatography. The calibration by NIR spectroscopy for evaluate the oil content of sunflower seeds, was determined by linear regression with the parameters estimated by partial least squares. The goodness of fit was evaluated by the coefficient of determination (R2), standard error of calibration (SEC) and the standard error of prediction (SEP). The R2 was 0.87, the SEC was 2.39, and the SEP was 1.97. Furthermore, the oil contents from 19 hybrids analysed by the NIR spectroscopy, but not included in the calibration, were similar to the chemical analysis. These similarities and the values of R2, SEC and SEP allowed us to infer that the analysis of intact seeds using the NIR spectroscopy is appropriate to discriminate the oil content of sunflower genotypes cultivated in Brazil.