Abstract
New Zealand (NZ) is the world’s largest exporter of dairy products. NZ milk products are known for their ‘clean and green’ reputation with high nutritional properties. In fact, the quality of NZ milk is perceived to be one of the best milk in the world. Cow, goat, and sheep milk are produced in NZ. In fact, NZ goat milk (GM) and sheep milk (SM) are considered as high-value products. The price of goat and sheep milk is 2-3x the price of cow milk (CM) in NZ. These facts put these high-value dairy products at risks of fraudulent activities like milk adulteration or counterfeit. Milk can be adulterated with different adulterants including water, vegetable protein, whey, and milk from another species. In other cases, expensive milk (e.g., GM, SM) could be adulterated with cheap milk (CM) for the purpose of economic gain. This is also known as economically motivated adulteration (EMA). It is therefore important to develop quick, effective, and robust tools for the detection of adulteration. Such techniques must be robust and high throughput requires a small amount of sample, and highly reproducible. Of particular interest is the application of proton nuclear magnetic resonance (1H-NMR) fingerprinting technique, which meets most of the mentioned requirements.
To date, there is limited information on the metabolite composition of milk produced in NZ. Moreover, NZ milk are not well studied in terms of their compositional properties. Therefore, this thesis is aimed to explore the capability for NMR-based metabolomics technique in detection of adulteration of NZ GM and SM with different concentrations of CM. To achieve this, the study was split into two parts. In the first part, NMR spectroscopy was used to characterise NZ CM, GM, and SM powder to select the discriminant metabolites for each species. In the second part, NMR was used to detect adulteration of GM and SM with different concentrations of CM. Advanced chemometrics (supervised and unsupervised approaches) were applied for data interpretation. SPSS and R studio were also employed for statistical analysis.
Overall, NMR fingerprinting technique alongside advanced chemometrics enabled detection of 17, 24, and 23 metabolites present in the water-soluble fractions of CM, GM, and SM, respectively. Out of the identified metabolites, carbohydrates, carboxylic acid, and amino acid were amongst the selected discriminant compounds in CM. In GM, the selected discriminant compounds include amino acid, fatty acid, nucleosides, carbohydrates, and carboxylic acid. Lastly, compounds such as carboxylic acid, carbohydrate, and nucleotide were selected as discriminant markers of SM. Following characterization, NMR spectroscopy was also successful in identifying potential markers of adulteration in GM and SM with CM. Based on VID feature selection procedure and Tukey’s test, phosphocholine was selected as a candidate marker of adulteration of GM with CM. On the other hand, N-acetyl carbohydrates and orotate can be proposed as potential markers of adulteration of SM with CM.
This work is the first study to characterize NZ milk types (CM, GM, SM) using NMR-based metabolomics, and attempt to detect adulteration of NZ GM and SM with different concentrations of CM. Overall, the NMR-fingerprinting technique was successful in characterising the metabolites present in the different milk types and detecting adulterations. Advanced chemometrics (supervised and unsupervised approach) were also suitable for the interpretation of NMR data, and for identifying discriminants, and in detecting adulterants. Further investigation of different milk fractions (such as lipid) and also the use of other fingerprinting techniques (e.g., LC-Q-TOF-MS, infrared) is needed to support the findings of the present study.
Keywords: NMR, metabolomics, chemometrics, cow milk, goat milk, sheep milk, adulteration
detection, metabolites, markers