The authentication of food products from the presence of non-allowed components for certain religion like lard is very important. In this study, we used proton Nuclear Magnetic Resonance ((1)H-NMR) spectroscopy for the analysis of butter adulterated with lard by simultaneously quantification of all proton bearing compounds, and consequently all relevant sample classes. Since the spectra obtained were too complex to be analyzed visually by the naked eyes, the classification of spectra was carried out.The multivariate calibration of partial least square (PLS) regression was used for modelling the relationship between actual value of lard and predicted value. The model yielded a highest regression coefficient (R(2)) of 0.998 and the lowest root mean square error calibration (RMSEC) of 0.0091% and root mean square error prediction (RMSEP) of 0.0090, respectively. Cross validation testing evaluates the predictive power of the model. PLS model was shown as good models as the intercept of R(2)Y and Q(2)Y were 0.0853 and -0.309, respectively.
Curcuma longa, Curcuma xanthorrhiza, and Curcuma manga have been widely used for herbal or traditional medicine purposes. It was reported that turmeric plants provided several biological activities such as antioxidant, anti-inflammatory, hepatoprotector, cardioprotector, and anticancer activities. Authentication of the Curcuma species is important to ensure its authenticity and to avoid adulteration practices. Plants from different origins will have different metabolite compositions because metabolites are affected by soil nutrition, climate, temperature, and humidity. 1H-NMR spectroscopy, principal component analysis (PCA), and orthogonal projections to latent structures-discriminant analysis (OPLS-DA) were used for authentication of C. longa, C. xanthorrhiza, and C. manga from seven different origins in Indonesia. From the 1H-NMR analysis it was obtained that 14 metabolites were responsible for generating classification model such as curcumin, demethoxycurcumin, alanine, methionine, threonine, lysine, alpha-glucose, beta-glucose, sucrose, alpha-fructose, beta-fructose, fumaric acid, tyrosine, and formate. Both PCA and OPLS-DA model demonstrated goodness of fit (R2 value more than 0.8) and good predictivity (Q2 value more than 0.45). All OPLS-DA models were validated by assessing the permutation test results with high value of original R2 and Q2. It can be concluded that metabolite fingerprinting using 1H-NMR spectroscopy and chemometrics provide a powerful tool for authentication of herbal and medicinal plants.