Rapid screening of tuna samples for food safety issues related to histamine content using fourier-transform mid-infrared (FT-MIR) and chemometrics

Abstract

Biogenic amines (BAs) generally result from the decarboxylation reaction of free amino acids as a result of the activity of different microorganisms. A build-up of these compounds can result in food being spoilt. Therefore, the rapid and precise detection of BAs like histamine is an important task for food safety. This research aimed to explore the potential of Fourier-Transform Mid-Infrared (FT-MIR) spectroscopy combined with chemometric methods to assess histamine in fresh tuna quantitatively. Based on the FT-MIR data, partial least squares regression models for the prediction of histamine were successfully constructed with R2 > 0.90. Machine learning algorithms (partial least squares-discrimination analysis, k-nearest neighbors, and support vector machine) were applied, and excellent discrimination results were achieved based on the limits specified in two different legislations (EU and FDA). The results support the use of a rapid, economic and reliable approach for the discrimination of samples that could pose a health risk to consumers.