ABSTRACT

This study was conducted to characterize the potential of near-infrared reflectance spectroscopy (NIRS) for the analysis of seed quality traits in intact, single seeds of rapeseed. Single seeds were scanned by NIRS using a special adapter and further analysed by the corresponding reference methods. Oil content was determined by solvent extraction in a set of 125 seeds. The fatty acid composition of the seed oil was analysed by gas chromatography in a set of 605 seeds. Protein content was determined by the combustion method in a set of 143 seeds. Glucosinolate content was analysed by high-performance liquid chromatography in a set of 145 seeds. Reliable calibration equations were developed for oil (r² = 0.87 in crossvalidation), protein (r² = 0.91), and glucosinolate content (r² = 0.90), oleic acid (r² = 0.89) and erucic acid concentration (r² = 0.90). The standard errors of cross validation were of 1.9% for oil content (mean ± SD of the calibration set = 44.3 ± 5.3%), 0.94% for protein content (21.1 ± 3.0%), 10.3 µmol g^-1 for glucosinolate content (51.2 ± 32.8 µmol g^-1), 8.3% for oleic acid (40.9 ± 25.1%), and 6.3% for erucic acid (21.2 ± 20.1%). No reliable calibration equations were obtained for the other fatty acids. The results of the present study demonstrate the practicability of NIRS for screening single seeds of rapeseed for quality traits.

INTRODUCTION

Breeding Brassica oil crops for modified fatty acid composition of the seed oil has been based on single-seed selections by gas chromatographic analyses of half seeds (Röbbelen, 1990). The selection efficiency for other traits such as oil, protein or glucosinolate content on a single seed basis has not been determined. In maize, Silvela et al. (1989) have demonstrated that a procedure based on selection between single kernels within an ear is more efficient to increase oil content than a procedure based on selection between plants in a row.

Nowadays near infrared reflectance spectroscopy (NIRS) is widely used for the analysis of bulk rapeseed samples for oil, protein and glucosinolate content, and the fatty acid composition of the seed oil. Analyses on small samples of about 60 mg intact seeds have been reported (Velasco and Becker, 1998). A further improvement in the application of NIRS technique for rapeseed breeding would imply the reduction of the sample size to a single seed. NIRS has been applied to single-seed analyses of oil and protein content in corn and soybean (Orman & Schumann, 1992; Dyer & Feng, 1995; Abe et al., 1995), protein content in wheat (Abe et al., 1995), oil content in meadowfoam (Limnanthes spp.; Patrick & Jolliff, 1997), oleic and linoleic acid concentrations in sunflower (Sato et al., 1995; Velasco et al., 1999a), oil content, oleic, linoleic, and erucic acid concentrations in rapeseed (Sato et al., 1998; Velasco et al., 1999b). The objective of this work was to study the potential of NIRS to estimate the oil, protein, and glucosinolate content, and the fatty acid composition of the seed oil in intact, single seeds of rapeseed.

MATERIALS AND METHODS

The seeds used for this study were chosen from a wide range of breeding materials of Brassica napus L.. The corresponding plants were grown during the years 1994 to 1997, either in the field or in greenhouses. The calibration sets consisted of 125 seeds for oil content, 143 seeds for protein content, 145 seeds for glucosinolate content, and 605 seeds for concentrations of individual fatty acids.

The intact seeds were first analysed by NIRS and then by the corresponding reference method. For NIRS analyses, the seeds were placed in an special adapter about 2-mm thick, with a total diameter of 37 mm and a central hole of 3 mm diameter. The adapter was inserted in a NIRS standard ring cup and the absorbance spectra (log 1/R) from 400 to 2500 nm were recorded at 2 nm intervals on a monochromator NIR Systems model 6500. Original reflectance spectra were corrected prior to calibration by applying first or second derivative transformation, standard normal variate, and de-trend scatter correction. Calibration equations were developed by using modified partial least squares (MPLS) regression.

Oil content (% fresh seed weight) was determined by repeated extraction with a solution i-propanol:petrolether (2:3). The fatty acid composition of the seed oil (% of the total fatty acids) was analysed by gas liquid chromatography (GLC) of fatty acid methyl esters. Protein content (% fresh seed weight) was analysed by the Dumas combustion method. Glucosinolate content (% fresh seed weight) was determined by high-performance liquid chromatography (HPLC) of desulphoglucosinolates.

RESULTS AND DISCUSSION

Calibration equations for oil, protein, and glucosinolate contents, and oleic and erucic acid concentrations in the seed oil showed high coefficients of determination in cross-validation, from 0.87 to 0.91 (Table 1). The ratios of the standard error of cross-validation (SECV) to the standard deviation (SD) of the reference values were between 0.30 and 0.36, demonstrating the usefulness of the calibration equations developed in the present study for single-seed selection.

TABLE 1. Calibration and cross validation statistics for analysis of oil content (% of fresh seed weight), protein content (% of fresh seed weight), glucosinolate (GSL) content (µmol g^-1 fresh seed weight), and major fatty acids (% of the total fatty acids) in intact, single seeds of rapeseed by near infrared reflectance spectroscopy (NIRS).

† SEC=standard error of calibration.

‡ SECV=standard error of cross validation.

The calibration plots for the five traits included in Table 1 are shown in Fig. 1 to Fig. 4, where the close relationship between NIRS and reference values can be appreciated. Calibration equations for the other major fatty acids were also developed, although the calibration and validation statistics did not reach the good values obtained for oleic and erucic acid. The ratio SECV/SD was of 0.65 for linoleic and of 0.70 for linolenic acid (data not shown), compared with a ratio of 0.33 and 0.31 for oleic and erucic acid, respectively.

Fig. 1 Calibration plot for oil content (% of fresh seed weight) in intact, single seeds of rapeseed. R²=coefficient of multiple determination in calibration, SECV=standard error of cross validation, SD=standard deviation of reference values.

Fig. 2 Calibration plot for protein content (% of fresh seed weight) in intact, single seeds of rapeseed. R²=coefficient of multiple determination in calibration, SECV=standard error of cross validation, SD=standard deviation of reference values.

The NIRS calibration equations developed in this study permit the simultaneous analysis of the oil, protein and glucosinolate contents, and the oleic and erucic acid concentrations of the seed oil in a single, intact seed of rapeseed in a nondestructive, fast, cost-effective and reliable way. Previous studies reported the evaluation of oil content and concentrations of some individual fatty acids in intact rapeseed (Sato et al., 1998; Velasco et al., 1999b). In the present study, we have demonstrated that other traits currently analysed by NIRS in bulk rapeseed samples, such as the protein and glucosinolate contents, can also be measured in single seeds.

The possibility of using NIRS for measuring a wide range of quality traits in single seeds opens up a new way in rapeseed breeding for seed quality. The usefulness of single-seed selection for the concentration of individual fatty acids is well known by rapeseed breeders, who currently use chromatographic analyses of half seeds. The use of NIRS offers clear advantages over gas chromatography for large screenings. It still has to be demonstrated the efficiency of single-seed selection for traits subjected to maternal effects, such as oil, protein, and glucosinolate contents, but a nondestructive analytical method for them is already available according to the results of the present study.

Fig. 3 Calibration plot for total glucosinolate content (µmol g-1 of fresh seed weight) in intact, single seeds of rapeseed. R²=coefficient of multiple determination in calibration, SECV=standard error of cross validation, SD=standard deviation of reference values.

Fig. 4 Calibration plots for oleic and erucic acid concentrations (% of the total fatty acids) in intact, single seeds of rapeseed. R²=coefficient of multiple determination in calibration, SECV=standard error of cross validation, SD=standard deviation of reference values.

REFERENCES

1. Abe, H., T. Kusama, S. Kawano, and M. Iwamoto, 1995. Non-destructive determination of protein content in a single kernel of wheat and soybean by near infrared spectroscopy. In: A.M.C. Davies & P. Williams (Eds.), Near Infrared Spectroscopy: the Future Waves, pp. 457-461. NIR Publications, Chichester, UK.

2. Dyer, D.J., and P. Feng, 1995. Near infrared applications in the development of genetically altered grains. In: A.M.C. Davies & P. Williams (Eds.), Near Infrared Spectroscopy: the Future Waves, pp. 490-493. NIR Publications, Chichester, UK.

3. Orman, B.A., and R.A. Schumann, Jr., 1992. Nondestructive single-kernel oil determination of maize by near-infrared transmission spectroscopy. Journal of the American Oil Chemists’ Society 69:1036-1038.

4. Patrick, B.E., and G.D. Jolliff, 1997. Nondestructive single-seed oil determination of meadowfoam by near-infrared transmission spectroscopy. Journal of the American Oil Chemists’ Society 74:273-276.

5. Röbbelen, G., 1990. Mutation breeding for quality improvement. A case study for oilseed crops. Mutation Breeding Review, n. 6, FAO/IAEA Division of Nuclear Techniques in Food and Agriculture. Vienna, Austria.

6. Sato, T., Y. Takahata, T. Noda, T. Yanagisawa, T. Morishita, S. Sakai, 1995. Nondestructive determination of fatty acid composition of husked sunflower (Helianthus annuus L.) seeds by near-infrared spectroscopy. Journal of the American Oil Chemists’ Society 72:1177-1183.

7. Sato, T., I. Uezono, T. Morishita, and T. Tetsuka, 1998. Nondestructive estimation of fatty acid composition in seeds of Brassica napus L. by near-infrared spectroscopy. Journal of the American Oil Chemists’ Society 75:1877-1881.

8. Silvela, L., R. Rodgers, A. Barrera, and D.E. Alexander, 1989. Effect of selection intensity and population size on percent oil in maize, Zea mays L. Theoretical and Applied Genetics 78:298-304.

9. Velasco, L., and H.C. Becker, 1998. Estimating the fatty acid composition of the oil in intact-seed rapeseed (Brassica napus L.) by near-infrared reflectance spectroscopy. Euphytica 101:221-230.

10. Velasco, L., B. Pérez-Vich, and J.M. Fernández-Martínez, 1999a. Nondestructive screening for oleic and linoleic acid in single sunflower achenes by near-infrared reflectance spectroscopy. Crop Science 39:219-222.

11. Velasco, L., C. Möllers, and H.C. Becker, 1999b. Estimation of seed weight, oil content and fatty acid composition in intact single seeds of rapeseed (Brassica napus L.) by near infrared reflectance spectroscopy. Euphytica 106:79-85.