电子舌结合气相色谱-质谱法分析评价不同莲蓉馅料风味差异

Analysis and evaluation of flavor differences in different lotus seed paste fillings by electronic tongue combined with gas chromatography-mass spectrometry

中文关键词:  莲蓉馅料  电子舌测定  游离脂肪酸含量  风味评价预测模型

英文关键词:lotus seed paste fillings  electronic tongue measurement  free fatty acid content  flavor evaluation prediction model

中文摘要:

      目的 利用电子舌和气相色谱-质谱(gas chromatography-mass spectrometry, GC-MS)技术检测纯白莲蓉、纯红莲蓉、大师手工纯白莲蓉、低糖纯白莲蓉和绿色食品级纯白莲蓉馅料5种不同品种的莲蓉馅料的风味差异。方法 采用电子舌对莲蓉馅料的甜味等主要风味指标进行评价, 结合GC-MS技术分析脂肪酸成分, 并通过偏最小二乘法回归分析(partial least squares regression, PLSR)建立电子舌和游离脂肪酸的评价模型, 预测莲蓉馅料的风味。结果 5种莲蓉馅料在甜味和咸味上存在显著差异(P<0.05), 其中, 绿色食品级纯白莲蓉馅料的甜味强度最高, 手工纯白莲蓉馅料的咸味强度最高。利用GC-MS技术共检测出油酸和亚油酸等9种游离脂肪酸, 在电子舌和脂肪酸联合评价的PLSR回归模型中, 甜味、鲜味、丰富性和油脂味模型的决定系数R2在校正集上分别为: 0.94、0.94、0.95、0.91, 在交叉验证集上分别为: 0.90、0.88、0.91、0.83, 其均方根误差(root mean squared error, RMSE)在校正集上分别为: 0.21、0.14、0.14、0.18, 在交叉验证集上分别为0.28、0.21、0.20、0.25。与莲蓉馅料单独电子舌或游离脂肪酸风味评价相比, 联合评价表现出更好的预测能力, 特别在油脂味评价上, 交叉验证集的决定系数大幅度提升, 从电子舌评价的0.15、游离脂肪酸评价的0.67增加到0.83, 决定系数更接近于1, 均方根误差较小, 显示出更高的拟合度。结论 电子舌和脂肪酸联合评价为莲蓉馅料风味的仪器评价研究提供参考数据。

英文摘要:

      Objective To investigate the flavor differences of 5 different varieties of lotus seed paste fillings, including pure white, pure red, master’s handmade pure white, low-sugar pure white, and green food-grade pure white by electronic tongue and gas chromatography-mass spectrometry (GC-MS) techniques. Methods The main flavor attributes, such as sweetness, of the lotus seed paste fillings were evaluated using an electronic tongue. GC-MS analysis was used to analyze the fatty acid composition. Partial least squares regression (PLSR) was employed to establish evaluation models combining electronic tongue and free fatty acids for predicting the flavor of the lotus seed paste fillings. Results Significant differences were observed among 5 varieties of lotus seed paste fillings in terms of sweetness and saltiness (P<0.05). Among them, the green food-grade pure white filling showed the highest sweetness intensity, while the master handcrafted pure white filling exhibited the highest saltiness intensity. A total of nine free fatty acids, including oleic acid and linoleic acid, were detected using GC-MS. In the PLSR regression model combining electronic tongue and fatty acids, the determination coefficients (R2) for sweetness, umami, richness, and oiliness models were 0.94, 0.94, 0.95 and 0.91, respectively, on the calibration set, and 0.90, 0.88, 0.91 and 0.83, respectively, on the cross-validation set. The root mean squared errors (RMSE) were 0.21, 0.14, 0.14 and 0.18 on the calibration set, and 0.28, 0.21, 0.20 and 0.25 on the cross-validation set. Compared to the separate evaluation of lotus seed paste fillings using electronic tongue or free fatty acids, the combined evaluation displayed improved predictive capability, particularly for oiliness evaluation. The determination coefficient on the cross-validation set increased significantly from 0.15 in the electronic tongue evaluation and 0.67 in the evaluation of free fatty acids to 0.83, approaching unity, with a smaller RMSE, indicating a higher degree of fit. Conclusion The combined evaluation of electronic tongue and fatty acids provides reference data for the instrumental evaluation of the flavor of lotus fillings.