Machine Learning Approaches Applied to GC-FID Fatty Acid Profiles to Discriminate Wild from Farmed Salmon

Latitudinal changes in blubber fatty acids of the South American sea lion along the Southeast Pacific Coast

The composition of marine mammal blubber is influenced by factors such as diet and environmental conditions. Here, we investigate the thermal influence on fatty acid composition and degree of unsaturation in the blubber of 151 South American sea lions (Otaria byronia). Samples were collected at ten locations along the Chilean coast and spanned a latitudinal range of approximately 2500 km, an arc of 23°. We found a significant latitudinal gradient in the biochemical composition of sea lion blubber. At higher colder latitudes the endogenous fatty acids in the blubber were more unsaturated, as indicated by the desaturation index. In cold conditions the modification of fatty acids from saturated to unsaturated improves fluidity of cell membranes, which is important for thermal insulation. Thus, the higher degree of unsaturated endogenous fatty acids at higher latitudes may help sea lions maintain thermal balance in colder environments. For the dietary-sourced fatty acids however, we found the opposite latitudinal pattern, here the overall degree of unsaturation decreases with latitude. The lower proportion of polyunsaturated fatty acids in blubber at higher latitudes may be influenced by the sea lion's dietary choices. Additionally, the presence of extremely high levels of 18:2n-6 in some sea lions south of 40°S, values more typical of terrestrial origin, suggests they consumed farmed salmon. The observed variation in fatty acid profiles across the latitudinal gradient implies that differences in fatty acid composition are not solely attributed to dietary variation, but potentially also to endogenous metabolism in response to environmental conditions.

MALDI-ToF MS and chemometric analysis as a tool for identifying wild and farmed salmon

In this study, the difference between wild and farmed salmon production was successfully profiled and differentiated by matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-ToF MS) combined with chemometric analysis. The established method based on multivariate analysis mainly involved principal component analysis (PCA), partial least squares-discriminant analysis (PLS-DA), and orthogonal partial least squares-discriminant analysis (OPLS-DA) as the screening and verifying tools to provide insights into the distinctive features found in wild and farmed salmon products, respectively. The discrimination between farmed and wild salmon was accomplished with 100% classification accuracy using chemometric models, 100% identification accuracy was also achieved in distinguishing wild Salmo salar and Oncorhynchus nerka samples. The results of the present work suggest that the proposed method could serve as a reference for detecting salmon fraud relating to wild or farmed production and expand the application of MALDI-ToF technology further into food authenticity applications.

Machine Learning Approach to Comparing Fatty Acid Profiles of Common Food Products Sold on Romanian Market

Food composition issues represent an increasing concern nowadays, in the context of diverse food commodity varieties. The contents and types of fatty acids are a constant preoccupation among consumers because of their reflections of nutrition and health problems. This study aims to find the best tool for the rapid and reliable identification of similarities and differences among several food items from a fatty acid profile perspective. An acknowledged GC-FID method was considered, while, for a better interpretation of the analytical results, machine learning algorithms were used. It was possible to develop a recognition model able to simultaneously differentiate, with an accuracy of 79.3%, nine product types using the bagged tree ensemble model. The low number of samples or some similarities among the classes could be responsible for the wrong assignments that occurred, especially in the biscuit, wafer and instant soup classes. Better accuracies values of 95, 86.1, and 97.8% were obtained when the products were grouped into three categories: (1) sunflower oil, mayonnaise, margarine, and cream cheese; (2) biscuits, cookies, margarine, and wafers; and (3) sunflower oil, chips, and instant soup.

Determination of Fatty Acid Profile in Processed Fish and Shellfish Foods

Seafood products are a crucial dietary source of n-3 polyunsaturated fatty acids (n-3 PUFA), which are essential for human health. However, the presence of these n-3 PUFA may be subject to changes related to different processing methods. The aim of this study was to determine the fatty acid composition, focusing on n-3 PUFA, in different processed fish and shellfish products of both EU and non-EU origin. The products were purchased from supermarkets and ethnic food shops in Messina (Italy). Gas chromatography with a flame-ionization detector (GC-FID) was used for analysis. Based on the fatty acid profile, the atherogenicity index (AI), thrombogenicity index (TI), and flesh lipid quality index (FLQ) were determined: 0.13-1.04 (AI), 0.19-0.89 (TI), and 0.41-29.90 (FLQ). The percentages of saturated (SFA), monounsaturated (MUFA), and polyunsaturated (PUFA) fatty acids fell within the following ranges: 13.55-50.48%, 18.91-65.58%, and 13.84-52.73%, respectively. Considering that all samples showed low AI and TI indices and that all processed fish products proved to be a good source of beneficial PUFAs, especially eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), their consumption is recommended for humans.

Hybrid volatilomics in cancer diagnosis by HS-GC-FID fingerprinting

Assessing volatile organic compounds (VOCs) as cancer signatures is one of the most promising techniques toward developing non-invasive, simple, and affordable diagnosis. Here, we have evaluated the feasibility of employing static headspace extraction (HS) followed by gas chromatography with flame ionization detector (GC-FID) as a screening tool to discriminate between cancer patients (head and neck-HNC,n= 15; and gastrointestinal cancer-GIC,n= 19) and healthy controls (n= 37) on the basis of a non-target (fingerprinting) analysis of oral fluid and urine. We evaluated the discrimination considering a single bodily fluid and adopting the hybrid approach, in which the oral fluid and urinary VOCs profiles were combined through data fusion. We used supervised orthogonal partial least squares discriminant analysis for classification, and we assessed the prediction power of the models by analyzing the values of goodness of prediction (Q²Y), area under the curve (AUC), sensitivity, and specificity. The individual models HNC urine, HNC oral fluid, and GIC oral fluid successfully discriminated between healthy controls and positive samples (Q²Y = 0.560, 0.525, and 0.559; AUC = 0.814, 0.850, and 0.926; sensitivity = 84.8, 70.2, and 78.6%; and specificity = 82.3; 81.5; 87.5%, respectively), whereas GIC urine was not adequate (Q²Y = 0.292, AUC = 0.694, sensitivity = 66.1%, and specificity = 77.0%). Compared to the respective individual models, Q²Y for the hybrid models increased (0.623 for hybrid HNC and 0.562 for hybrid GIC). However, sensitivity was higher for HNC urine and GIC oral fluid than for hybrid HNC (75.6%) and hybrid GIC (69.8%), respectively. These results suggested that HS-GC-FID fingerprinting is suitable and holds great potential for cancer screening. Additionally, the hybrid approach tends to increase the predictive power if the individual models present suitable quality parameter values. Otherwise, it is more advantageous to use a single body fluid for analysis.

The use of various statistical methods for authenticity and detection of adulteration in fish and seafood

Various methodologies including genetic analyses, morphometrics, proteomics, lipidomics, metabolomics, etc. are now used or being developed to authenticate fish and seafood. Such techniques usually lead to the generation of enormous amounts of data. The analysis and interpretation of this information can be particularly challenging. Statistical techniques are therefore commonly used to assist in analyzing these data, visualizing trends and differences and extracting conclusions. This review article aims at presenting and discussing statistical methods used in studies on fish and seafood authenticity and adulteration, allowing researchers to consider their options based on previous successes/failures but also offering some recommendations about the future of such techniques. Techniques such as PCA, AMOVA and FST statistics, that allow the differentiation of genetic groups, or techniques such as MANOVA that allow large data sets of morphometric characteristics or elemental differences to be analyzed are discussed. Furthermore, methods such as cluster analysis, DFA, CVA, CDA and heatmaps/Circos plots that allow samples to be differentiated based on their geographical origin are also reviewed and their advantages and disadvantages as found in past studies are given. Finally, mathematical simulations and modeling are presented in a detailed review of studies using them, together with their advantages and limitations.

Discrimination between Wild and Farmed Sea Bass by Using New Spectrometry and Spectroscopy Methods

European sea bass (Dicentrarchus labrax L.) is one of the most economically important fish species in the Mediterranean Sea area. Despite strict requirements regarding indications of production method (wild/farmed), incorrect labelling of sea bass is a practice still frequently detected. The aim of this study was to evaluate the capabilities of two techniques, Near-InfraRed (NIR) spectroscopy and mass spectrometry, to discriminate sea bass according to the production method. Two categories were discriminated based on the docosahexaenoic and arachidonic fatty acid ratio by using a Direct Sample Analysis (DSA) system integrated with a time-of-flight (TOF) mass spectrometer. The cut-off value of 3.42, of fatty acid ratio, was able to discriminate between the two types of fish with sensitivity and specificity of 100%. It was possible to classify fish production by using multivariate analysis with portable NIR. The results achieved by the developed validation models suggest that this approach is able to distinguish the two product categories with high sensitivity (100%) and specificity (90%). The results obtained from this study highlight the potential application of two easy, fast, and accurate screening methods to detect fraud in commercial sea bass production.

A Brief Review of Machine Learning-Based Bioactive Compound Research

Bioactive compounds are often used as initial substances for many therapeutic agents. In recent years, both theoretical and practical innovations in hardware-assisted and fast-evolving machine learning (ML) have made it possible to identify desired bioactive compounds in chemical spaces, such as those in natural products (NPs). This review introduces how machine learning approaches can be used for the identification and evaluation of bioactive compounds. It also provides an overview of recent research trends in machine learning-based prediction and the evaluation of bioactive compounds by listing real-world examples along with various input data. In addition, several ML-based approaches to identify specific bioactive compounds for cardiovascular and metabolic diseases are described. Overall, these approaches are important for the discovery of novel bioactive compounds and provide new insights into the machine learning basis for various traditional applications of bioactive compound-related research.

Fingerprint Approaches Coupled with Chemometrics to Discriminate Geographic Origin of Imported Salmon in China's Consumer Market

Of the salmon sold in China’s consumer market, 92% was labelled as Norwegian salmon, but was in fact was mainly imported from Chile. The aim of this study was to establish an effective method for discriminating the geographic origin of imported salmon using two fingerprint approaches, Near-infrared (NIR) spectroscopy and mineral element fingerprint (MEF). In total, 80 salmon (40 from Norway and 40 from Chile) were tested, and data generated by NIR and MEF were analysed via various chemometrics. Four spectral preprocessing methods, including vector normalization (VN), Savitzky Golay (SG) smoothing, first derivative (FD) and second derivative (SD), were employed on the raw NIR data, and a partial least squares (PLS) model based on the FD + SG9 pretreatment could successfully differentiate Norwegian salmons from Chilean salmons, with a R² value of 98.5%. Analysis of variance (ANOVA) and multiple comparative analysis were employed on the contents of 16 mineral elements including Pb, Fe, Cu, Zn, Al, Sr, Ni, As, Cr, V, Se, Mn, K, Ca, Na and Mg. The results showed that Fe, Zn, Al, Ni, As, Cr, V, Se, Ca and Na could be used as characteristic elements to discriminate the geographical origin of the imported salmon, and the discrimination rate of the linear discriminant analysis (LDA) model, trained on the above 10 elements, could reach up to 98.8%. The results demonstrate that both NIR and MEF could be effective tools for the rapid discrimination of geographic origin of imported salmon in China’s consumer market.

Changes of Volatile Flavor Compounds in Large Yellow Croaker (Larimichthys crocea) during Storage, as Evaluated by Headspace Gas Chromatography-Ion Mobility Spectrometry and Principal Component Analysis

The large yellow croaker is one of the most economically important fish in Zhoushan, Zhejiang Province, and is well known for its high protein and fat contents, fresh and tender meat, and soft taste. However, the mechanisms involved in its flavor changes during storage have yet to be revealed, although lipid oxidation has been considered to be one important process in determining such changes. Thus, to explore the changes in the flavor of large yellow croaker fish meat during different storage periods, the main physical and chemical characteristics of the fish meat, including the acid value, peroxide value, p-anisidine value, conjugated diene value, and identities of the various flavor substances, were investigated and analyzed by multivariable methods, including headspace gas chromatography-ion mobility spectrometry (GC-IMS) and principal component analysis (PCA). It was found that after 60 d storage, the types and contents of the aldehyde and ketone aroma components increased significantly, while after 120 d, the contents of ketones (2-butanone), alcohols (1-propanethiol), and aldehydes (n-nonanal) decreased significantly. More specifically, aldehyde components dominated over ketones and lipids, while the n-nonanal content showed a downward trend during storage, and the 3-methylbutanol (trimer), 3-methylbutanol (dimer, D), 3-pentanone (D), and 3-pentanone (monomer) contents increased, whereas these compounds were identified as the key components affecting the fish meat flavor. Furthermore, after 120 d storage, the number of different flavor components reached its highest value, thereby confirming that the storage time influences the flavor of large yellow croaker fish. In this context, it should be noted that many of these compounds form through the Maillard reaction to accelerate the deterioration of fish meat. It was also found that after storage for 120 d, the physical indices of large yellow croaker meat showed significant changes, and its physicochemical properties varied. These results therefore demonstrate that a combination of GC-IMS and PCA can be used to identify the differences in flavor components present in fish meat during storage. Our study provides useful knowledge for understanding the different flavors associated with fish meat products during and following storage.