Kim Yookyung, Himmelsbach David S, Kays Sandra E
Agricultural Research Service, U.S. Department of Agriculture, Athens, Georgia 30604-5677, USA.
J Agric Food Chem. 2007 May 30;55(11):4327-33. doi: 10.1021/jf063729l. Epub 2007 May 2.
Fourier transform mid-infrared (FT-IR) spectroscopy was investigated as a method of analysis for trans fatty acid content of cereal products without the need for prior oil extraction. Spectra were obtained, with an FT-IR spectrometer equipped with an attenuated total reflectance (ATR) device, of ground samples pressed onto the diamond ATR surface, and trans fatty acids were measured by a modification of AOAC Method 996.01. Partial least-squares (PLS) models were developed for the prediction of trans fatty acids in ground samples using several wavenumber selections on the basis of bands related to lipids. The models (n = 79) predicted trans fatty acids in ground samples with standard error of cross-validation (SECV) of 1.10-1.25 (range 0-12.4) % and R2 of 0.85-0.88 and in validation samples (n = 26) with standard error of performance (SEP) of 0.96-1.12 (range 0-12.2) % and r2 of 0.89-0.92, indicating sufficient accuracy for screening. Sample trans fatty acid % was predicted as accurately with the fingerprint region (1500-900 cm(-1)) as with the entire range (4000-650 cm(-1)) indicating, in concert with the regression coefficients, the importance of the isolated trans double bonds at 966 cm(-1) in development of the model. Data is also presented on prediction of trans fatty acids using the spectra of residual oil films on the ATR surface after removing the solid portion of the sample.
傅里叶变换中红外（FT-IR）光谱法被作为一种无需预先提取油脂就能分析谷物产品反式脂肪酸含量的方法进行了研究。使用配备衰减全反射（ATR）装置的FT-IR光谱仪，对压在金刚石ATR表面的磨碎样品进行光谱采集，并且通过对AOAC方法996.01进行修改来测定反式脂肪酸。基于与脂质相关的谱带，利用几种波数选择，建立了偏最小二乘（PLS）模型来预测磨碎样品中的反式脂肪酸。这些模型（n = 79）对磨碎样品中反式脂肪酸的预测，交叉验证标准误差（SECV）为1.10 - 1.25（范围0 - 12.4）%，R2为0.85 - 0.88；在验证样品（n = 26）中，性能标准误差（SEP）为0.96 - 1.12（范围0 - 12.2）%，r2为0.89 - 0.92，表明具有足够的筛选准确性。样品反式脂肪酸百分比在指纹区（1500 - 900 cm(-1)）和整个范围（4000 - 650 cm(-1)）预测的准确性相同，这与回归系数一致，表明在模型建立中966 cm(-1)处孤立的反式双键的重要性。还给出了利用去除样品固体部分后ATR表面残留油膜的光谱预测反式脂肪酸的数据。
