Prediction of Aerobic Plate Count on Beef Surface Using Fluorescence Fingerprint

• Published in: Food and bioprocess technology 2014-05, Vol.7 (5), p.1496-1504
• Main Authors: Yoshimura, Masatoshi, Sugiyama, Junichi, Tsuta, Mizuki, Fujita, Kaori, Shibata, Mario, Kokawa, Mito, Oshita, Seiichi, Oto, Naomi
• Format: Article
• Language: English
• Subjects: Aerobic bacteria; Agriculture; bacterial contamination; Beef; Beef cattle; Biotechnology; Cattle; Chemical compounds; Chemistry; Chemistry and Materials Science; Chemistry/Food Science; data collection; Datasets; Fingerprints; flavins; Flora; Fluorescence; Fluorophores; Food irradiation; Food Science; lean meat; Meat; metabolism; Multivariate analysis; NAD; NAD (coenzyme); Nondestructive testing; Original Paper; plate count; Pollution load; Porphyrins; prediction; Predictions; Regression analysis; Regression coefficients; slicing; Spectrophotometry; Spectroscopy; Spectrum analysis; Tryptophan; Vitamin A; wavelengths
• ISSN: 1935-5130; 1935-5149
• DOI: 10.1007/s11947-013-1167-8

The potential of fluorescence fingerprint (FF) spectroscopy was investigated to develop a nondestructive prediction method of aerobic plate count on a beef surface. Sixty samples (e.g., 30 lean meat slices each of Australian cattle and Japanese cattle) stored aerobically at 15 °C were analyzed by front-face fluorescence spectrophotometry. FF and aerobic plate count (APC) were measured after 0, 12, 24, 36, and 48 h of storage. FFs were collected in both excitation and emission wavelength ranges of 200–900 nm. Partial least-squares regression (PLSR) performed on an FF dataset predicted an APC in the bacterial contamination load range from 1.7 to 7.8 log colony-forming units (cfu)/cm² with a prediction error of 0.752 log cfu/cm². The regions where the regression coefficient of the PLSR model was relatively high were consistent with those of the FF peaks of five intrinsic fluorophores: tryptophan, NAD(P)H, vitamin A, porphyrins, and flavins. This suggests that changes in the autofluorescence of these intrinsic fluorophores due to the metabolism of bacterial flora on meat are reflected in the PLSR model for predicting APC from the FF dataset. FF spectroscopy coupled with multivariate analysis appeared to be applicable to the nondestructive determination of APC on the surface of lean beef.