基于小鼠粪便气味信息的牛乳清蛋白功能效果的鉴别区分

以电子鼻气敏传感器阵列跟踪牛乳清蛋白干预小鼠粪便气味,以期建立基于气味信息无损评价牛乳清蛋白功能的新方法。动物实验结果表明,与NS组比较,给药牛乳清蛋白试验组血清中的超氧化物歧化酶(SOD)活性上升,丙二醛(MDA)含量显著下降,牛乳清蛋白在小鼠血清中具有抗氧化的能力。电子鼻实验单因素方差分析发现,样品量、顶空生成时间、顶空生成体积和载气流速对传感器S5影响均不显著,对其他传感器响应信号的影响显著。结合典则判别分析获得电子鼻检测较佳条件为:样品量1粒,顶空生成时间10 min,顶空生成体积150 mL,载气流速200 mL·min^-1。结合主成分分析和典则判别分析,电子鼻基本可将牛乳清蛋白不同灌胃剂量与对照组小鼠粪便样品区分,并区分不同灌胃剂量的不同周期组小鼠粪便样品。采用多元线性回归分析建立了小鼠体重的预测(R²=0.122 7),效果尚可。典型相关分析结果表明,粪便电子鼻气味信息与其生理指标两者有着极好的相关性,关联性较强。基于干预小鼠粪便气味信息,可以实现牛乳清蛋白干预与对照药物处理的区分,为食品体内功能性无创评价提供了方法思路。

Identification of functional effects of bovine whey protein based on odor information of mouse feces

In order to establish a new method for nondestructive evaluation of bovine whey protein function based on odor information, the gas sensor array of electronic nose (E-nose) was used to track the flavor of fecal samples of mice affected by bovine whey. The results showed that compared with the NS group, the superoxide dismutase (SOD) activity in the serum of the bovine whey protein test group was increased, and the malondialdehyde (MDA) content was significantly decreased. The bovine whey protein has antioxidant capacity in the mouse serum. The electronic nose experiment showed that the effects of sample weight, headspace generation time, headspace generation volume and flow rate carrier gas on the response signal of the sensor were studied by one-way ANOVA. It was found that the response of E-nose sensors was affected by sample weight, headspace generation time, headspace generation volume and flow rate carrier gas significantly, except for S5. The experiment condition for E-nose was optimized as followed: 1 sample, 10 min of headspace generation time, 150 mL of headspace generation volume and 200 mL·min^-1 of carrier gas flow rate. To distinguish the difference of doses of bovine whey protein and control group, principle component analysis (PCA) and canonical discriminant analysis (CDA) could discriminate the samples of different groups, and could distinguish the fecal samples of mice of different cycle groups with different gavage doses. For the prediction of the weight, the multiple linear regression analysis (MLR) was employed, and effect predictive model was built for weight (R²=0.122 7), the effect was acceptable. The typical correlation analysis results showed that the odor information of electronic stool had an excellent correlation with its physiological indicators, and the correlation was strong. Based on the odor of fecal samples of mice, it’s possible to distinguish the effect of bovine whey protein treatment with the controls, providing an idea for the non-invasive evaluation of functional food in vivo.