Novel application of artificial senses (E-Nose and E-Tongue) and chemometrics approach for rapid organoleptic evaluation of Milk

Flavour is composed principally of the sensations of smell and taste. To separate and identify sapid substances, traditional methods of chemistry have been employed. By now, methods such as these have been applied and whole books have been devoted to the sensory assessment of milk. The origin of the significant flavorants in milk probably originate from the cow's metabolism; however, there is potential for manipulating flavour by means of feed. The aim of this study was to evaluate the efficiency of artificial senses combined with chemometric technique used almost like a human panel for to evaluate the effect of olive cake supplementation on the organoleptic/aromatic profile of cow's milk, in a process of valorization of by-products in the circular economy. In fact, the use of olive cake as animal feed represents a sustainable alternative reducing the costs associated with animal nutrition. For the trial, 24 multiparous dairy cows, breeding in the same condition, were divided into two homogeneous groups: Control (C), received conventional diet; Treated (T) received a conventional diet integrated with 7% of dried olive cake. Raw milk collection took place in the spring period (April-May 2022), in 3 sampling dates: after 15 days of dietary adaptation, then at 20 and 40 days of the trial. In the laboratory, the individual samples of milk from each group were mixed and analyzed 10 times using an E-nose (FOX 4000, Alpha M.O.S., Toulouse, France) with 18 MOS sensors and a potentiometric E-tongue (aAstree, Alpha M.O.S., Toulouse, France) with 7 chemical sensors. The results were subjected to exploratory PCA and expressed based upon the discrimination index (DI). Then, for increase the probability of correct classification, a low-level data fusion was investigated and the single-matrix with data from all sensors was subjected to a new PCA. On the basis of Mahalanobis distance, the pattern discrimination index (PDI) among the 4 groups was calculated.PCA analysis of raw data was able to discriminate all sampling groups, highlighting the effect of both time and diet. The DI was 83% and the first two components (PC1 and PC2), represented 95% of the total variance between the sample measurement. Observing the PCA plot, the groups are distributed along the first component (PC1 = 78.68%) as a function of the diet, while in upwards direction, along the second component (PC2 = 19.79%) as a function of time. The PDI results confirmed that the greatest differences in the taste and smell of the milk from the two groups are due to time. The organoleptic distances between the diet related samples are very small, confirming the capacity of E-nose and E-tongue to find even minimal differences. The results highlighted by the artificial senses, shows that is possibleto suggest the use of E-nose and E-tongue as effective tools to recognize a product made with alternative eco-friendly resources, addressed to minimize the environmental impact of the olive oil industry