Süt ve Süt Ürünlerinde Tür Tayinin Senkronize Floresans Spektroskopisiyle Belirlenmesi

Abstract

In the dairy industry, the mixing of high-priced milk types such as sheep, goats, and buffalo with low-priced milk types such as cow milk, and therefore the inability to determine the purity of milk types and the fact that true labeling cannot be done correctly have become a common problem worldwide. Recently developed spectroscopic methods for solving these problems are preferred because they are faster, easier, and less costly than traditional methods based on DNA and protein. In the scope of the thesis, synchronous fluorescence spectroscopy (SFS) method was developed in order to identify cows, sheep, goats, and buffalo species in raw milk and fermented milk products (yogurt and cheese). The principle of the method is based on differences in the types and amounts of fluorophore compounds found in milk and dairy products. In the study, qualitative and quantitative analyzes were performed for each sample group (milk, yogurt, and cheese) with data recorded in the excitation wavelength range of 250-550 nm with Δλ = 20-100 nm in 10 nm steps. For qualitative analysis; Three models have been developed based on the smallest square-discriminant analysis (PLS-DA) method for the separation of pure and mixture samples, the classification of four different milk types and the definition of binary mix types (sheep-cow, goat-cow, and buffalo-cow). In quantitative analysis, after determining the binary mix types with the PLS-DA model, three partial smallest square (PLS) models were developed to determine the mixing ratios of cow milk in other milk types. With the PLS-DA models developed for the raw milk sample group, all milk types were successfully separated from each other. When the performances of PLS models of sheep-cow, goat-cow and buffalo-cow binary mixtures obtained by mixing with cow milk are examined, the values of high calibration and validation determination coefficient (R2) (sheep: 0.992-0.989; goat: 0.992-0.988; buffalo: 0.998-0.986) with low limit of detection (LOD) and limit of quantification (LOQ) (sheep: 4.06%-12.31%; goat: 3.52%-10.66%; buffalo: 3.52%-10.64%) values were obtained. In 3 different PLS models developed to determine the mixing ratios of binary milk mixtures in yogurt samples, R2 values of high calibration and validation (sheep-cow: 0.996-0.957; goat-cow: 0.992-0.978; buffalo-cow: 0.999-0.978) with low LOD and LOQ ( sheep-cow: 1.64% -5.47%; goat-cow: 3.25% - 10.85%; buffalo-cow: 0.46% -5.5%) values show a successful performance. When the PLS models of binary mixtures in cheese samples are examined, it has been shown that the milk ratios contained in the mixtures can be successfully determined with high calibration and validation R2 values (sheep-cow: 0.998-0.969; goat-cow: 0.996-0.981; buffalo-cow: 0.981-0.952) and low LOD and LOQ values (sheep-cow: 1.47%-4.90%; goat-cow: 3.28%-10.94%; buffalo-cow: 1.73%-5.75%).