摘要：Stress triggers a battery of physiological responses in fish, including the activation of metabolic pathways involved in energy production, which helps the animal to cope with the adverse situation. Prolonged exposure to stressful farming conditions may induce adverse effects at the whole-animal level, impairing welfare. Fourier transform infrared (FTIR) spectroscopy is a rapid biochemical fingerprinting technique, that, combined with chemometrics, was applied to disclose the metabolic alterations in the fish liver as a result of exposure to standard stressful practices in aquaculture. Gilthead seabream (Sparus aurata) adults exposed to different stressors were used as model species. Spectra were preprocessed before multivariate statistical analysis. Principal components analysis (PCA) was used for pattern recognition and identification of the most discriminatory wavenumbers. Key spectral features were selected and used for classification using the k-nearest neighbour (KNN) algorithm to evaluate whether the spectral changes allowed for the reliable discrimination between experimental groups. PCA loadings suggested that major variations in the hepatic infrared spectra responsible for the discrimination between the experimental groups were due to differences in the intensity of absorption bands associated with proteins, lipids and carbohydrates. This broad-range technique can thus be useful in an exploratory approach before any targeted analysis.
其他摘要：Abstract Stress triggers a battery of physiological responses in fish, including the activation of metabolic pathways involved in energy production, which helps the animal to cope with the adverse situation. Prolonged exposure to stressful farming conditions may induce adverse effects at the whole-animal level, impairing welfare. Fourier transform infrared (FTIR) spectroscopy is a rapid biochemical fingerprinting technique, that, combined with chemometrics, was applied to disclose the metabolic alterations in the fish liver as a result of exposure to standard stressful practices in aquaculture. Gilthead seabream ( Sparus aurata ) adults exposed to different stressors were used as model species. Spectra were preprocessed before multivariate statistical analysis. Principal components analysis (PCA) was used for pattern recognition and identification of the most discriminatory wavenumbers. Key spectral features were selected and used for classification using the k-nearest neighbour (KNN) algorithm to evaluate whether the spectral changes allowed for the reliable discrimination between experimental groups. PCA loadings suggested that major variations in the hepatic infrared spectra responsible for the discrimination between the experimental groups were due to differences in the intensity of absorption bands associated with proteins, lipids and carbohydrates. This broad-range technique can thus be useful in an exploratory approach before any targeted analysis.
