Clara Perez-Gonzalez

The quality of milk obtained from cows is influenced by a variety of factors, one of the most Fifteen significant being the composition of the feed intake. This study evaluates the impact of different feeding regimens on raw milk composition by combining conventional chemical analysis with a potentiometric bioelectronic tongue (bioET) specifically designed for raw milk analysis, incorporating lipase-, β-galactosidase-, and galactose-oxidase-based membranes. Milk samples from cows fed a standard diet (SF) were compared with those from cows supplemented with essential fatty acids (EFAF), tannins and plant-derived fatty acids (TA), or 3-nitrooxypropanol (3-NOP). Multivariate analysis of conventional chemical data revealed that EFAF and TA supplementation significantly affected fat-related parameters, including total fatty acids (TFA), saturated fatty acids (SFA), and short-chain fatty acids (SCFA), whereas 3-NOP did not produce detectable changes, in agreement with its known metabolic mechanism. Classification of samples according to feeding regimen using Support Vector Machines with a radial basis function (SVM-RBF) kernel yielded classification errors of 21.1% (EFAF) and 24.6% (TA). The bioET results confirmed these trends, showing improved classification performance (17.1% for EFAF and 20.3% for TA) and no significant effect for 3-NOP. In addition, the bioET exhibited strong correlations with fat-related parameters as well as lactose, consistent with the enzymatic specificity of the sensor array. High prediction accuracy was achieved for key variables such as fat (R² = 0.96) and SFA (R² = 0.97), enabling reliable estimation of multiple components from a single rapid measurement. These findings highlight the potential of the optimized bioelectronic tongue as a rapid, non-invasive tool for detecting diet-induced changes in raw milk and for predicting major compositional parameters, offering a promising approach for nutritional monitoring in dairy production.

The adulteration of milk presents significant challenges in the food industry, promoting the need for efficient detection methods. This study introduces a potentiometric electronic tongue for rapid and accurate milk adulteration detection. Utilizing polymeric membranes integrated with various additives, the electronic tongue distinguishes between different milk types and detects common adulterants. Experimental results demonstrate its effectiveness in discriminating raw, pasteurized, and medicated cow milk, as well as goat milk. Moreover, it successfully identifies adulterants like water and bovine milk in goat milk samples. Chemometric analyses, including Principal Component Analysis and Partial Least Squares regression, correlate sensor responses with traditional milk parameters such as fat, protein, and lactose content with up to a 0.97 R² on the validation step. Strong correlations validate the electronic tongue's potential for rapid milk quality assessment. This innovative approach offers a cost-effective, reliable solution for detecting milk adulteration in contrast with current techniques that require numerous, time consuming experiments.