Citation: FU Dan-Dan , WANG Qiao-Hua , GAO Sheng , MA Mei-Hu . Analysis of S-Ovalbumin Content of Different Varieties of Eggs during Storage and Its Nondestructive Testing Model by Visible-Near Infrared Spectroscopy. Chinese Journal of Analytical Chemistry, 2020, 48(2): 289-297. doi: 10.19756/j.issn.0253-3820.191329 [复制]
Analysis of S-Ovalbumin Content of Different Varieties of Eggs during Storage and Its Nondestructive Testing Model by Visible-Near Infrared Spectroscopy
利用可见/近红外光纤光谱采集罗曼粉壳和海蓝褐壳两个品种的鸡蛋在349-1000 nm的透射光谱，对270枚鸡蛋的天然卵白蛋白的S-型空间构象异构体（S-卵白蛋白，S-ovalbumin，S-ova）含量进行了定量分析，实现了不同品种鸡蛋中S-卵白蛋白含量的快速无损检测。通过比较贮期不同品种鸡蛋的平均光谱发现，两个品种鸡蛋的光谱吸收峰位置相同，仅可见光范围内的光谱吸收能量值有所不同。通过标准正态变量校正（SNV）对原始光谱进行预处理，并利用无信息变量消除算法（UVE）从500~950 nm的全光谱中提取了67个特征波长，建立的偏最小二乘（PLS）回归模型可以很好地预测不同品种的S-卵白蛋白含量。为了更进一步消除特征波长之间的多重共线性，利用逐步回归（Stepwise regression）算法对特征波长进行二次筛选，最终筛选出了16个特征波长，建立多元回归模型，其校正集的决定系数（R2）为0.9511，均方根误差（RMSE）为0.0478，预测集的R2为0.8380，RMSE为0.1116，预测集相对分析误差（RPD）为2.2620。此模型对预测集中50个罗曼粉壳鸡蛋和40个海蓝褐壳鸡蛋样本的R2分别为0.8119和0.9116，RMSE分别为0.1298和0.0834，模型适用性更佳。本研究结果表明，可见/近红外光谱能够对不同品种的S-卵白蛋白含量进行无损检测，建立的通用预测模型为开发便携式蛋白含量无损检测装置奠定了基础。
The visible-near-infrared (Vis-NIR) transmission spectroscopy technique was used to analyze the content of S-ovalbumin (S-ova), which had high correlation with egg freshness, and to establish a nondestructive prediction model. The visible/near-infrared fiber spectroscopy were used to collect the transmission spectrum of two varieties of eggs at 349-1000 nm, and the S-ovalbumin content of 270 eggs was measured by wet chemistry method. By comparing the average spectra of eggs of different varieties during storage, it was found that the spectral absorption peaks of different varieties of eggs had the same position, and only the spectral energy values in the visible range differed. The original spectrum was preprocessed by standard normal variate (SNV), and 67 characteristic wavelengths were extracted from the full spectrum of 500-950 nm using uninformative variables elimination (UVE). It was concluded that partial least squares (PLS) regression model based on 67 characteristic wavelengths could predict the S-ovalbumin content. To further eliminate the multi-collinearity between the characteristic wavelengths, the stepwise regression algorithm was used to perform secondary screening on the characteristic wavelengths, and finally 16 characteristic wavelengths were selected. By using the 16 characteristic wavelengths to establish a multivariate regression model, the coefficient of determination (R2) of the training set was 0.9511, the root mean square error (RMSE) was 0.0478, and the R2 of the prediction set was 0.8380. Besides, the RMSE was 0.1116, and the residual predictive deviation (RPD) was 2.2620. The general predictive model was used to predict the S-ovalbumin content of 50 eggs with Roman pink shell and 40 eggs with sea blue brown shell in the prediction set. The R2 of the predicted and measured values were 0.8119 and 0.9116, respectively, and the RMSEs were 0.1298 and 0.0834, respectively. Therefore, the general model could perform nondestructive testing on the S-ovalbumin content of these two different varieties of eggs better, and the model was more applicable. The results showed that the visible/near-infrared spectroscopy could accurately detect the S-ovalbumin content of eggs in different varieties, and the established general prediction model laid a foundation for the development of portable non-destructive testing device for protein content.
Hirose M. Foods Food Ingred. J. JP.,2005,210(8):770-777
Huang Q, Qiu N, Ma M H. Poultry Sci.,2012,91(3):739-743
Fu D D, Wang Q H, Ma M H, Ma Y X. Int. J. Food Propert.,2019,22(1):1077-1086
Nawar S, Mouazen A M. Soil Tillage Res.,2019,190:120-127
Zhu Z H, Li W Q, Wang Q H, Tang Y. J. Food Process Engineer.,2017, 40(3):UNSP12345
Dong X G, Dong J, Peng Y K, Tang X Y. Spectrosc. Lett.,2017,50(9):463-469
Syduzzaman M, Rahman A, Alin K, Fujitani S, Kashimori A, Suzuki T, Ogawa Y, Kondo N. Engineer. Agric. Environ. Food, 2019,12(3):289-296
Dong J, Dong X G, Li Y L, Peng Y K, Chao K L, Gao C Y, Tang X Y. Comput. Electron. Agric.,2019,157:471-478[LM] 14 WU Jian-Hu, HUANG Jun. Modern Food Science & Technology,2015, (5):285-290
Suktanarak S, Teerachaichayut S. J. Food Engineer.,2017,215:97-103.
Feng C H, Makino Y, Yoshimura M, Francisco J. Rodríguez P. Food Chem.,2018,264:419-426
Qi H J, Tarin P K, Arnon K, Jin X, Li S W. Soil Tillage Res.,2018,175:267-275