Perilla, an oilseed crop grown in Asian countries, is consumed in oil form as a spice or medicine and in its leafy form as a vegetable. The seed is made up of several fatty acids, especially linolenic acid, which has been found to be beneficial for controlling chronic disease. Breeding the seeds to produce a crop with an extremely high fatty acid composition makes them more marketable.The raw NIR reflectance (a) and the second-derivative spectra (b) of intact perilla seed samples are shown. Images reprinted with permission of the Journal of Agricultural and Food Chemistry.In the past, the fatty acid content was determined by gas-liquid chromatography and flame ionization detection with a capillary column. These methods required a multistep preparation process that included converting fatty acids to fatty acid methyl esters for gas-liquid chromatography analysis. Although these methods were very accurate, they were time-consuming, labor-intensive and expensive, and the seeds had to be pulverized, so they were destroyed in the process.A group from Mokpo National University in Muan and from Kangwon National University in Samcheok, both in South Korea, has developed a near-infrared reflectance spectroscopy method to determine the oil content in intact perilla seeds.The researchers gathered 397 seed samples of perilla germ plasm — the DNA organism used to breed the species — to develop the near-infrared spectroscopy model. Through a process of homogenization, extraction and drying of the germ plasm, oil content was determined as a percentage of the dry seed weight. The scientists analyzed the fatty acids using an NIRSystem model 6500 near-infrared scanning monochromator from Foss NIRSystems Inc. of Silver Spring, Md. The scanning was completed in 1.5 minutes per sample. The laboratory reference values are shown compared with values predicted by near-infrared reflectance spectroscopy (NIRS) in the validation set for oil content and oleic acid (a), linolenic acid (b) and oil content (c) of intact perilla seeds. (GLC = gas-liquid chromatography, SEP = standard error of prediction.) One challenge they faced was in determining which wavelength or region in the near-infrared spectrum was closest to the contents of the natural compounds they were analyzing. According to researcher Si Hyung Park, the investigators used sophisticated statistical processes to find the specific wavelengths that correlate to the samples. The method is fast and nondestructive, and it can be performed while studying other components such as protein, water and functional compounds, which are all important in determining nutritional quality to select superior breeding lines.The next step is to update and improve the equations used with future samples from different environments and germ plasms, as well as to improve calibration models by studying a broader range of oil and fatty acid values of samples. By advancing this method of near-infrared reflectance spectroscopy to determine the oil content and fatty acid composition of perilla seeds, this staple of Asian diets can be bred for its beneficial components.Journal of Agricultural and Food Chemistry, March 7, 2007, pp. 1679-1685.Contact: Si Hyung Park, Mokpo National University, South Korea; email:shp@mokpo.ac.kr.