Key Laboratory of Geriatric Nutrition and Health, Ministry of Education, Beijing Technology & Business University, Beijing 100048, China.
Key Laboratory of Geriatric Nutrition and Health, Ministry of Education, Beijing Technology & Business University, Beijing 100048, China.
Food Res Int. 2024 Sep;191:114705. doi: 10.1016/j.foodres.2024.114705. Epub 2024 Jun 27.
Ultra-high temperature (UHT) milk is popular among consumers. However, its flavor and texture change in its shelf life. Flavor is highly determinative for the success of dairy products and for consumers' willingness to buy. It is important to milk producers to ensure the optimal flavor of their products in the shelf life. In order to be able to control and predict the flavor quality of UHT milk during the shelf life, this study compared the variations in sensory quality, volatile aroma release and backbone flavor factors and developed a discriminant model to assess flavor quality based on flavouromics data of five competing milk sample during storage. Using partial least squares discriminant analysis (PLS-DA) with Electronic-nose (E-nose) data excellent classification sensitivity and specificity were achieved compared to models based on gas chromatography-mass spectrometry (GC-MS) data. The PLS-DA model using E-nose data exhibited a 100% correct classification rate for the storage period, and a 92% correct rate based on the eight variable importance in the projection (VIP) elements screened for volatile components from different groups. The discriminative model developed herein based on E-nose combined with chemometrics demonstrated advantages such as speed, efficiency, and environmental friendliness. This method shows promise as a precise tool for analyzing aroma changes in UHT milk during its shelf life, and provide support for controlling the flavor substances and milk product development.
超高温（UHT）牛奶在消费者中很受欢迎。然而，它的风味和质地在保质期内会发生变化。风味对乳制品的成功和消费者的购买意愿有很大的决定性。对于牛奶生产商来说，确保其产品在保质期内具有最佳的风味非常重要。为了能够在保质期内控制和预测 UHT 牛奶的风味质量，本研究比较了感官质量、挥发性香气释放和骨干风味因素的变化，并开发了一种基于五种竞争牛奶样品在储存过程中风味组学数据的判别模型来评估风味质量。与基于气相色谱-质谱（GC-MS）数据的模型相比，使用电子鼻（E-nose）数据的偏最小二乘判别分析（PLS-DA）模型实现了出色的分类灵敏度和特异性。使用 E-nose 数据的 PLS-DA 模型对储存期的正确分类率为 100%，对来自不同组的挥发性成分筛选的 8 个投影变量重要性（VIP）元素的正确分类率为 92%。本文基于 E-nose 结合化学计量学开发的判别模型具有速度快、效率高、环保等优点。该方法有望成为分析 UHT 牛奶在保质期内香气变化的精确工具，并为控制风味物质和牛奶产品开发提供支持。
