Multivariate statistical analysis of stable isotope signatures and element concentrations to differentiate the geographical origin of retail milk sold in Singapore

Singapore's reliance on imported foods makes the country susceptible to food fraud and food safety related incidents. To ensure authenticity of food source and indirectly food quality and safety, the development of an independent system that can trace the origin of a food is important to compliment the existing paper-based approach. A scientific approach to food traceability using isotope ratio and element concentration measurements was demonstrated using dairy milk as an example matrix. An iterative stepwise linear discriminant analysis procedure applied to the isotope ratios and element concentrations of 57 milk samples separated them into their six different country of origins. Si, Se, Li, B, Rb, Ba, P, Mn, Mo, Pb and the δ13C and δ15N of the defatted milk casein were found to be the key parameters providing maximum discrimination. The model was successfully evaluated with a blind-test, comprising of 10 samples with origins unknown to the analysts. •307 milk samples analysed for their isotope signatures and element profiles.•Geographical origin of milk samples could be definitively differentiated.•Iterative stepwise linear discriminant analysis can help to improve discrimination.