FT-IR analysis of tilapia fillets: Developing PLS models for the prediction of storage days, aerobic plate count, and lactobacilli

Abstract

Chemometric models using FT-IR and Partial Least Squares (PLS) regression were developed for the rapid determination of aerobic plate count (APC), lactobacilli count, and storage time of tilapia fillets. A comprehensive data set of 137 spectra was obtained from sampled tilapia over a 10-month period. IR absorption bands of methyl and methylene stretching vibrations, amide I and II, and 1500–1025 cm−1 regions provided the greatest variance for the models. Successful discrimination of stored tilapia within a 2-day range and accurate prediction of APC and lactobacilli counts with RMSE values lower than ± 1 Log10 CFU·g−1 were achieved.