Prediction of amino acids in freeze dried pork by near infrared reflectance spectroscopy

Abstract

NIRS was used to predict the amino acid profile of freeze-dried pork samples. Samples (150; Longissimus thoracis et lumborum) of pork were used for analysis. After freeze drying, samples were analyzed using HPLC to find out the amino acid content. Samples were scanned and partial least squares (PLS) regression methods were used to predict the amino acid. The determination coefficient obtained by full cross-validated (80 as a sample for calibration set, 25 samples as a validation set) PLS models indicated that the NIR original spectra had an excellent ability to predict the contents of alanine, proline and methionine. Prediction of glutamic acid and glycine using standard normalized variate (SNV) pretreatment of spectral modeling was accurate. Similarly, prediction of arginine,tyrosine, valine, isoleucine, leucine, phenylalanine and lysine were accurate using SNV or multiplicative scattering correction (MSC) pre-processing spectra modeling. It was not possible to predict aspartic acid, serine, threonine, cystine, and histidine. These results indicated that the NIRS can be used for prediction of selected amino acids in the freeze dried pork.