| With the development of economy and the improvement of living standard,people pay more and more attention to food safety.However,due to the large population of our country,the demand for food is very high,and part of the farmers have a shallow sense of quality and safety,and lack of professional knowledge of poultry breeding.In order to pursue short-term economic benefits,the food safety problems of poultry meat are more and more frequent,such as the long term or excessive addition of antibiotics in the feed.As the most commonly eaten poultry meat,the issues of food safety of chicken and duck meat are also highly concerned.In this study,gold glue was used as the enhanced substrate,and different pretreatment methods,including adaptive iterative punitive least squares（air-PLS）,normalization,second derivative and baseline,were used.Principal component analysis（PCA）,support vector machine（SVM）or linear discriminant analysis（LDA）are used to identify sulfamethazine（SM-2）and sulfapyridine（SPD）residues in chicken and duck meat,or identify neomycin（NEO）and chloramphenicol（CAP）residues in chicken and duck meat.The main research contents are as follows:（1）In the process of identifying SM-2 and SPD residues in chicken,the characteristic peak at 1188 cm-1was used to identify whether there was SM-2 residues in chicken or not.The characteristic peak at 937 cm-1was used to identify whether there was SPD residues in chicken or not.The single factor method was used to optimize the test conditions and the optimal test conditions were obtained as follows:the addition amount of sodium citrate to prepare gold glue was 3.7 m L,the addition amount of gold glue was 500μL,the addition amount of NaCl solution was 100μL,and the adsorption time was 5 min.Under the optimal experimental conditions,the rapid identification of SM-2 and SPD residues in chicken was carried out,and PCA-SVM classification model was established.The classification accuracy of this model was 93.23%.The sensitivity and specificity of the model were 77.42～100%and 96～99.02%,respectively.The sensitivity of chicken with SM-2+SPD was highest（100%）,and that of chicken with SPD was lowest（77.42%）.The specificity of chicken with SPD was highest（99.02%）,and that of chicken with SM-2 was lowest（96%）.（2）In the process of identifying SM-2 and SPD residues in duck meat,557 and 573cm-1Raman peaks were used to identify whether there were SM-2 and SPD residues in duck meat or not,respectively.According to the intensities of the Raman characteristic peaks at 557 and 573 cm-1,the optimal test conditions were determined:the amount of sodium citrate added to prepare the gold glue was 3.7 m L,the addition amount of gold glue was 450μL,the addition amount of NaCl solution was 150μL,and the adsorption time was 5 min.After the original Raman spectrum was pretreated by air-PLS+second derivative,the classification accuracy of the model established by PCA combined with SVM was 90.44%,and the sensitivity and specificity of the classification model for four types of duck meat samples in the test set were 83.33～96.97%and 89.32～100%,respectively.The sensitivity of duck meat with SPD was the highest（96.97%）,and that of duck meat with SM-2 was lowest（83.33%）.The specificities of duck meat containing SM-2+SPD and blank duck meat were highest（100%）,and that of duck meat containing SPD was lowest（89.32%）.（3）The Raman spectra of NEO and CAP residues in chicken were studied.The characteristic peak at 546 cm-1was determined to identify whether there was NEO residues in chicken or not,and the characteristic peak at 666 cm-1was determined to identify whether there was CAP residues in chicken or not.After baseline correction+second derivative preprocessing,the classification model was established by PCA combined with LDA,and the classification accuracy of the model reached 100%,which could completely identify four types of chicken samples,and the sensitivity and specificity of the classification model for four types of chicken samples in the test set were 100%.（4）In the study of NEO and CAP residues in duck meat,the characteristic peak at 542cm-1was used to identify whether there was NEO residues in duck meat or not,and the characteristic peak at 735 cm-1was used to identify whether there was CAP residues in duck meat or not.According to the classification accuracy of the test set,the normalization+second derivative was selected as the spectral preprocessing method.The classification accuracy of the model established by PCA combined with LDA was99.24%.The sensitivity and specificity of the model for four kinds of duck meat samples in the test set were 96.97～100%and 98.99～100%,respectively.Among them,the sensitivity of blank duck meat samples was lowest（96.97%）,and the sensitivities of other three kinds of samples were all 100%.The specificity of duck meat samples with CAP was lowest,98.99%,and the specificities of other three types were all 100%.In conclusion,it was feasible to identify SM-2 and SPD residues in poultry meat and identify NEO and CAP residues in poultry meat by SERS and chemometrics. |
PDF Full Text Request