Discrimination of Cocoa Beans According to Geographical Origin by Electronic Tongue and Multivariate Algorithms

Metabolomics identify landscape of food sensory properties

Sensory evaluation is a key component of food production strategy. The classical food sensory evaluation method is time-consuming, laborious, costly, and highly subjective. Since flavor (taste and smell), texture, and mouthfeel are all related to the chemical properties of food, there has been a growing interest in how they affect the senses of food. In the past decades, emerging metabolomics has received much attention in the field of sensory evaluation, because it not only offers a broad picture of chemical composition for sensory properties but also revealed their changes and functions in food proceeding. This article reviewed food chemicals regarding the flavor, smell, and texture of foods, and discussed the advantages and limitations of applying metabolomics approaches to sensory evaluation, including GC-MS, LC-MS, and NMR. Taken together, this review gives a comprehensive, critical overview of the current state, future challenges, and trends in metabolomics on food sensory properties.

Multi-sensor integration approach based on hyperspectral imaging and electronic nose for quantitation of fat and peroxide value of pork meat

The study assessed the feasibility of merging data acquired from hyperspectral imaging (HSI) and electronic nose (e-nose) to develop a robust method for the rapid prediction of intramuscular fat (IMF) and peroxide value (PV) of pork meat affected by temperature and NaCl treatments. Multivariate calibration models for prediction of IMF and PV using median spectra features (MSF) and image texture features (ITF) from HSI data and mean signal values (MSV) from e-nose signals were established based on support vector machine regression (SVMR). Optimum wavelengths highly related to IMF and PV were selected from the MSF and ITF. Next, recurring optimum wavelengths from the two feature groups were manually obtained and merged to constitute "combined attribute features" (CAF) which yielded acceptable results with (R_c² = 0.877, 0.891; RMSEC = 2.410, 1.109; R_p² = 0.790, 0.858; RMSEP = 3.611, 2.013) respectively for IMF and PV. MSV yielded relatively low results with (R_c² = 0.783, 0.877; RMSEC = 4.591, 0.653; R_p² = 0.704, 0.797; RMSEP = 3.991, 0.760) respectively for IMF and PV. Finally, data fusion of CAF and MSV was performed which yielded relatively improved prediction results with (R_c² = 0.936, 0.955; RMSEC = 1.209, 0.997; R_p² = 0.895, 0.901; RMSEP = 2.099, 1.008) respectively for IMF and PV. The results obtained demonstrate that it is feasible to mutually integrate spectral and image features with volatile information to quantitatively monitor IMF and PV in processed pork meat. Graphical abstract.

Identification of geographical origin of Keemun black tea based on its volatile composition coupled with multivariate statistical analyses

BACKGROUND

Keemun black tea (KBT) is one of the most popular tea beverages in China as a result of its unique flavor and potential health benefits. The geographical origin of KBT influences its quality and price. The present study aimed to apply a head-space solid phase microextraction approach and gas chromatography-mass spectrometry combined with chemometric analysis to profile the volatile compounds of KBT collected from five production areas.

RESULTS

Thirty-one peaks were detected in 61 KBT samples. Hierarchical cluster analysis, principal component analysis (PCA), k-nearest neighbor (k-NN) and stepwise linear discriminant analysis (SLDA) were employed to visualize the volatile fractions. The results of unsupervised statistical tools were compared using a test for similarities and distinctions, which showed that different sources may be associated. A satisfying combination of average recognition (91.7%) and cross-validation prediction abilities (84.6%) was obtained for the PCA-k-NN. Among all of the statistical tools, SLDA provided promising results, with 100% recognition and 96.4% prediction ability.

CONCLUSION

The results obtained in the present study indicate that the volatile compounds can be used as indicators to identify the geographical origin of KBT. © 2019 Society of Chemical Industry.

Alternative data mining/machine learning methods for the analytical evaluation of food quality and authenticity - A review

In recent years, the variety and volume of data acquired by modern analytical instruments in order to conduct a better authentication of food has dramatically increased. Several pattern recognition tools have been developed to deal with the large volume and complexity of available trial data. The most widely used methods are principal component analysis (PCA), partial least squares-discriminant analysis (PLS-DA), soft independent modelling by class analogy (SIMCA), k-nearest neighbours (kNN), parallel factor analysis (PARAFAC), and multivariate curve resolution-alternating least squares (MCR-ALS). Nevertheless, there are alternative data treatment methods, such as support vector machine (SVM), classification and regression tree (CART) and random forest (RF), that show a great potential and more advantages compared to conventional ones. In this paper, we explain the background of these methods and review and discuss the reported studies in which these three methods have been applied in the area of food quality and authenticity. In addition, we clarify the technical terminology used in this particular area of research.

Taste Recognition in E-Tongue Using Local Discriminant Preservation Projection

Electronic tongue (E-Tongue), as a novel taste analysis tool, shows a promising perspective for taste recognition. In this paper, we constructed a voltammetric E-Tongue system and measured 13 different kinds of liquid samples, such as tea, wine, beverage, functional materials, etc. Owing to the noise of system and a variety of environmental conditions, the acquired E-Tongue data shows inseparable patterns. To this end, from the viewpoint of algorithm, we propose a local discriminant preservation projection (LDPP) model, an under-studied subspace learning algorithm, that concerns the local discrimination and neighborhood structure preservation. In contrast with other conventional subspace projection methods, LDPP has two merits. On one hand, with local discrimination it has a higher tolerance to abnormal data or outliers. On the other hand, it can project the data to a more separable space with local structure preservation. Further, support vector machine, extreme learning machine (ELM), and kernelized ELM (KELM) have been used as classifiers for taste recognition in E-Tongue. Experimental results demonstrate that the proposed E-Tongue is effective for multiple tastes recognition in both efficiency and effectiveness. Particularly, the proposed LDPP-based KELM classifier model achieves the best taste recognition performance of 98%. The developed benchmark data sets and codes will be released and downloaded in http://www.leizhang.tk/ tempcode.html.

System level analysis of cacao seed ripening reveals a sequential interplay of primary and secondary metabolism leading to polyphenol accumulation and preparation of stress resistance

Theobroma cacao and its popular product, chocolate, are attracting attention due to potential health benefits including antioxidative effects by polyphenols, anti-depressant effects by high serotonin levels, inhibition of platelet aggregation and prevention of obesity-dependent insulin resistance. The development of cacao seeds during fruit ripening is the most crucial process for the accumulation of these compounds. In this study, we analyzed the primary and the secondary metabolome as well as the proteome during Theobroma cacao cv. Forastero seed development by applying an integrative extraction protocol. The combination of multivariate statistics and mathematical modelling revealed a complex consecutive coordination of primary and secondary metabolism and corresponding pathways. Tricarboxylic acid (TCA) cycle and aromatic amino acid metabolism dominated during the early developmental stages (stages 1 and 2; cell division and expansion phase). This was accompanied with a significant shift of proteins from phenylpropanoid metabolism to flavonoid biosynthesis. At stage 3 (reserve accumulation phase), metabolism of sucrose switched from hydrolysis into raffinose synthesis. Lipids as well as proteins involved in lipid metabolism increased whereas amino acids and N-phenylpropenoyl amino acids decreased. Purine alkaloids, polyphenols, and raffinose as well as proteins involved in abiotic and biotic stress accumulated at stage 4 (maturation phase) endowing cacao seeds the characteristic astringent taste and resistance to stress. In summary, metabolic key points of cacao seed development comprise the sequential coordination of primary metabolites, phenylpropanoid, N-phenylpropenoyl amino acid, serotonin, lipid and polyphenol metabolism thereby covering the major compound classes involved in cacao aroma and health benefits.

Application of LogitBoost Classifier for Traceability Using SNP Chip Data

Consumer attention to food safety has increased rapidly due to animal-related diseases; therefore, it is important to identify their places of origin (POO) for safety purposes. However, only a few studies have addressed this issue and focused on machine learning-based approaches. In the present study, classification analyses were performed using a customized SNP chip for POO prediction. To accomplish this, 4,122 pigs originating from 104 farms were genotyped using the SNP chip. Several factors were considered to establish the best prediction model based on these data. We also assessed the applicability of the suggested model using a kinship coefficient-filtering approach. Our results showed that the LogitBoost-based prediction model outperformed other classifiers in terms of classification performance under most conditions. Specifically, a greater level of accuracy was observed when a higher kinship-based cutoff was employed. These results demonstrated the applicability of a machine learning-based approach using SNP chip data for practical traceability.