Abstract
There is a critical need for a rapid and simple method of qualitative and quantitative analysis of testosterone propionate (TP) and nandrolone (NT) residues in duck meat. In this study, we applied surface-enhanced Raman spectroscopy (SERS) coupled multivariate analysis for the classification and detection of TP and NT residues in duck meat. A total of 294 duck meat extract samples were obtained from duck breast meats based on a LC-MS/MS sample preparation method with slight modification including 102 duck meat extract samples without TP and NT, 43 duck meat samples containing TP, 47 duck meat extract samples containing NT, and 102 duck meat extract samples containing TP and NT. Raw Raman spectra were pretreated by using adaptive iteratively reweighted penalized least squares (airPLS), normalization and first derivative, and then the score values of first 10 principal components were selected as the inputs of the developed models. A particle swarm optimization-support vector classification (PSO-SVC) model was created to classify all the duck meat samples into the 4 groups (i.e., control group, TP group, NT group, and TP combined with NT group) with the classification accuracies of 99.49 and 100% for training set and test set, respectively. Furthermore, 2 least squares support vector regression (LS-SVR) models were developed to predict the TP values in samples with a determination coefficient (R(2)) value of 0.9316, root mean square error (RMSE) value of 2.1739, and ratio of prediction to deviation (RPD) value of 3.2189 for the test set, and NT values in samples with an R(2) value of 0.9038, RMSE value of 2.2914, and RPD value of 2.9701 for the test set. Surface-enhanced Raman spectroscopy technology, in combination with multivariate analysis, has the potential to become the qualitative and quantitative analysis tool for TP and NT residues in duck meat extract.