Accuracy of near infrared spectroscopy for prediction of chemical composition, salt content and free amino acids in dry-cured ham

The capability of near infrared (NIR) spectroscopy was examined for the purposes of quality control of the traditional Slovenian dry-cured ham "Kraski prsut." Predictive models were developed for moisture, salt, protein, non-protein nitrogen, intramuscular fat and free amino acids in biceps lemons muscle (n =135). The models' quality was assessed using statistical parameters: coefficient of determination (R-2) and standard error (se) of cross-validation (CV) and external validation (EV). Residual predictive deviation (RPD) was also assessed. Best results were obtained for salt content and salt percentage in moisture/dry matter (R-CV(2)> 0.90, RPD > 3.0), it was satisfactory for moisture, non-protein nitrogen, intramuscular fat and total free amino acids (R-CV(2)=0.75-0.90, RPD = 2.0-3.0), while not so for protein content and proteolysis index (R-CV(2) = 0.65-0.75, RPD < 2.0). Calibrations for individual free amino acids yielded R-CV(2) from 0.40 to 0.90 and RPD from 1.3 to 2.9. Additional external validation of models on independent samples yielded comparable results. Based on the results. NIR spectroscopy can replace chemical methods in quality control of dry-cured ham. (C) 2011 Elsevier Ltd. All rights reserved.