Search

The Atlantic cod was listed as ‘vulnerable’ by the International Union for Conservation of Nature, a condition that persists today. Fishing pressure on the Atlantic cod could be partially transferred to the Pacific cod, since the two cod species share genetic and phenotypic similarities. The aim of this study is to expand knowledge of the composition of dried salted cod obtained from Atlantic and Pacific cod species, with the Atlantic cod being from two different harvesting locations. The comparison of these cod species revealed the existence of nine significant differences among individual FAs (accountable for 63.2% of total FAs), which was at a similar level to that observed between different harvesting locations for the Atlantic cod (ten significant differences among individual FAs, accountable for 61.6% of total FAs). Canonical discriminant analysis and cross-validation achieved full discrimination of the cod’s origin and 100% accuracy in the cod’s origin classification. The amount of EPA plus DHA in dried salted cod reached its higher value among the Pacific cod (302.3 mg/100 g), while the Atlantic cod averaged 284.1 g/100 g of edible portion. The Pacific cod presented a higher α-tocopherol content than its Atlantic counterpart (8.04 vs. 4.94 µg/g).

Producción Científica, In the present study, the mineral composition of seventy-one bee pollen samples from four different apiaries was determined by means of inductively coupled plasma-optical emission spectrometry. The results showed that there were no significant differences in relation to the overall mineral content per sample in terms of the apiary of origin or the harvesting period; the most common elements were phosphorus and potassium with concentrations ranging from 2.3 to 5.1 g/kg (dry weight). Moreover, the mineral content of the samples analyzed was similar to or higher than the recommended as well-balanced food for bees. Regarding the nutritional value for humans, bee pollen samples could be considered as a food rich in copper, iron, magnesium, manganese, and phosphorus. Finally, a canonical discriminant analysis was performed, and it was found that the apiary of origin could be distinguished by using the first three canonical functions; furthermore, over 90% of the samples could be correctly assigned to their corresponding apiary. The results were even better as regards the harvesting period, as only the first two canonical functions were sufficient to make a distinction between the different harvesting periods, resulting in a perfect match (100% of success rate)., Ministerio de Economía y Competitividad (RTI2018-096268-B-100), Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria (RTA 2015-00013-C03-01 and 03), FEDER and FSE (project AGRISOST-CM S2018/BAA-4330)

Bee pollen is currently one of the most commonly consumed food supplements, as it is considered to be a good source of bioactive substances and energy. It contains various health-promoting compounds, such as proteins, amino acids, lipids, as well as glucosinolates. In the present study, the glucosinolate content was determined, by means of ultra-performance liquid chromatography coupled to a quadrupole time-of-flight mass detector, in 72 bee pollen samples from four different apiaries in Guadalajara (Spain), harvested in three different periods. In addition, 11 commercial multifloral samples from different Spanish regions were also analyzed. The aim was to verify the suitability of these compounds as biomarkers of their geographical origin, and to test their potential for distinguishing the harvesting period. By means of a canonical discriminant analysis, it was possible to differentiate the apiary of origin of most of the samples, and these could also be clearly differentiated from the commercial ones, simply as a result of the glucosinolate content. In addition, it was also demonstrated for the first time that bee pollen samples were capable of being differentiated according to the time of harvesting and their glucosinolate content.

Bee pollen may be contaminated with pyrrolizidine alkaloids (PAs) and their N-oxides (PANOs), which are mainly detected by liquid chromatography coupled to tandem mass spectrometry (LC–MS/MS), even though the use of fast near-infrared (NIR) spectroscopy is an ongoing alternative. Therefore, the main challenge of this study was to assess the feasibility of both a lab-stationary (Foss) and a portable (Polispec) NIR spectrometer in 60 dehydrated bee pollen samples. After an ANOVA-feature selection of the most informative NIR spectral data, canonical discriminant analysis (CDA) was performed to distinguish three quantitative PA/PANO classes (µg/kg): 400, high. According to the LC–MS/MS analysis, 77% of the samples were contaminated with PAs/PANOs and the sum content of the 17 target analytes was higher than 400 µg/kg in 28% of the samples. CDA was carried out on a pool of 18 (Foss) and 22 (Polispec) selected spectral variables and allowed accurate classification of samples from the low class as confirmed by the high values of Matthews correlation coefficient (≥ 0.91) for both NIR spectrometers. Leave-one-out cross-validation highlighted precise recognition of samples characterised by a high PA/PANO content with a low misclassification rate (0.02) as false negatives. The most informative wavelengths were within the 2400 nm regions for Foss and > 1500 nm for Polispec that could be associated with cyclic amines, and epoxide chemical structures of PAs/PANOs. In sum, both lab-stationary and portable NIR systems are reliable and fast techniques for detecting PA/PANO contamination in bee pollen.

The effect of storage time and packing method on dried Lycium fruits was studied through an electronic olfactory system with the metal oxide sensor array that provides an overall perception of the volatile compounds presented in the sample headspace. Principle component analysis (PCA), canonical discriminant analysis (CDA), and cluster analysis (CA) were used for freshness and packing methods discrimination of dried Lycium fruits. The stale samples of 2015 and 2016 could be separated with those of 2017 by PCA, CDA, and CA analysis. Better discrimination results were obtained by CDA, with samples of 2015 and 2016 separated with each other. For samples of 2017, the unpackaged samples of 2017-4 were distinguished with the vacuumed samples, while samples of grade C were separated with B and D. For quantitative analysis, predictive models for prediction of the storage years of dried Lycium fruits were built with methods of partial least square (PLS) analysis, multiple linear regression (MLR), and back propagation neural network (BPNN). The model built by BPNN showed the best predict ability with R2 = 0.9994, while PLS and MLR were also effective in the prediction of storage years of dried Lycium fruits, with high determination coefficients of 0.9316 and 0.9330. These findings showed that E-nose can be used in the discrimination of the storage time and package method of dried Lycium fruits.

The classification of six mushroom species (white beech, brown beech, button, oyster, king oyster, and enoki mushrooms) was successfully achieved using canonical discriminant analysis (CDA) on volatile metabolite data sets obtained by headspace-solid-phase microextraction gas chromatography (HS-SPME-GC). Twenty-seven major volatile compounds in six edible mushrooms were positively identified by HS-SPME-GC mass spectroscopy. The total volatile content was highest in brown beech mushroom (P < 0.05). Significant difference in volatile profile was observed between brown beach and white beech mushrooms. Button mushroom contained significantly higher contents of benzaldehyde and benzyl alcohol than the other mushrooms (P < 0.05). Oyster mushroom contained 1-octen-3-ol as the most prevalent volatile, representing 67% out of total volatiles. Hexanal (35.0%) and 1-octen-3-ol (22.5%) were the most abundant volatiles found in king oyster. Hexanal (29.1%) was the most prevalent volatile in enoki mushroom only. Several volatile pairs with very high positive correlation in their levels were identified, representing the highest correlation coefficient (r = 0.970) for the pair of t-2-octenal and 2,4-octandienal. CDA was much more efficient than principal component analysis for the differentiation of mushroom species. PRACTICAL APPLICATION: The present study provided the important information on the volatile metabolite profiles of popular six commercial mushroom species. The present data will be useful for the quality control of mushrooms cultivated in farms and mushroom products processed in food industry. The strategy of canonical discriminant analysis in combination with HS-SPME-GC could be expanded for the determining the authentication of mushroom species.

In Spain, ewe milk is mainly used for cheesemaking, and farming systems have traditionally been based on the use of autochthonous breeds. However, in recent years, the progressive introduction of highly productive foreign breeds in Spanish farms has led to an increasing interest in the characterization of dairy sheep breeds to evaluate whether genetic selection schemes should focus on productivity or milk technological aptitude. The purpose of this work was to explore milk composition and coagulation to classify 4 of the main dairy sheep breeds used in Spain. This study included 832 individual ewe milk samples from the breeds Manchega, Assaf, Merino de Grazalema, and Merino de Los Pedroches. Samples were analyzed for native pH, composition (fat, protein, lactose, and total solids), coagulation properties, and individual laboratory curd yield. An indicator of coagulation efficiency was also determined. Canonical discriminant analysis was performed to establish differences and similarities among breeds based on the measured variables. In addition, cluster analysis was performed to study and quantify the concrete relationships among the discriminated groups. Discriminant analysis proved to be a powerful tool to accurately draw distinctions between breeds. In all cases, discrimination among breeds was evident and the 4 breeds could be easily differentiated. Cluster analysis showed greater similarity between Merino de Grazalema and Assaf compared with the other breeds, and F-statistics indicated a higher discriminating ability for the variables related to milk composition. However, Merino de Grazalema and Manchega were difficult to separate according to milk composition, but the coagulation process differenced them clearly. Coagulation also evidenced similarities between Manchega and Merino de Los Pedroches, although the latter was revealed to be the most different breed of all 4, which could lay the ground for its differentiation as an independent breed in the Official Catalogue of Spanish Livestock Breeds.

Objectives Near-infrared (NIR) spectroscopy is a rapid technique able to assess meat quality even if its capability to determine the shelf life of chicken fresh cuts is still debated, especially for portable devices. The aim of the study was to compare bench-top and portable NIR instruments in discriminating between four chicken breast refrigeration times (RT), coupled with multivariate classifier models. Materials and Methods Ninety-six samples were analysed by both NIR tools at 2, 6, 10 and 14 days post mortem. NIR data were subsequently submitted to partial least squares discriminant analysis (PLS-DA) and canonical discriminant analysis (CDA). The latter was preceded by double feature selection based on Boruta and Stepwise procedures. Results PLS-DA sorted moderate separation of RT theses, while shelf life assessment was more accurate on application of Stepwise-CDA. Bench-top tool had better performance than portable one, probably because it captured more informative spectral data as shown by the variable importance in projection (VIP) and restricted pool of Stepwise-CDA predictive scores (SPS). Conclusions NIR tools coupled with a multivariate model provide deep insight into the physicochemical processes occurring during storage. Spectroscopy showed reliable effectiveness to recognise a 7-day shelf life threshold of breasts, suitable for routine at-line application for screening of meat quality.

The amount and quality of covering adipose tissue affect the suitability of hind legs for the production of high-quality seasoned hams. To date, no studies exist on the correlation between EUROP carcass classification and backfat fatty acid (FA) composition in heavy pigs used for dry-cured hams. A sample of 898 Italian Large White heavy pigs was used to verify the relationship between carcass classification based on lean meat percentage and backfat FA composition. A Canonical Discriminant Analysis (CDA) was used to verify the power of individual FA and FA categories in discriminating among EUROP classes. The results proved that saturated FAs (i.e. palmitic, stearic and arachidic acids) and the n-6 polyunsaturated FAs have the highest discriminating power, thus permitting to differentiate among E, U, R, O carcass classes. For the first time, this work demonstrates the relationship between EUROP pig carcass grading, which is only based on an estimate of the percentage of lean meat, and backfat FA composition.

Provenance of seafood items is a concern for consumers and importers alike. Elemental profiling has been put forth as a potential tool to improve seafood traceability. Whiteleg shrimp Litopenaeus vannamei collected from farms in Ecuador, India, Indonesia, Vietnam, and Thailand were classified to country of origin based on element concentrations in muscle tissue. Shrimp were dried, digested, and analyzed via ICP-MS. Out of the 41 elements investigated, 33 are reported as being above detection limits. Twenty-eight elements were statistically different across countries, and of these 28, Ecuador had unique group membership in 7. A random forest classification model utilizing 16 elements had an overall accuracy of 91% of correctly classified samples to country of origin. A canonical discriminant analysis was conducted to understand the variation in the data and identify elements that were important to differentiation in multi-dimension space. Elements identified as important contributors were Al, As, B, Ca, Co, Cs, Sr, and V. This study shows that shrimp from Ecuador tend to be more mineralized than shrimp from Asia, and classification models can discern samples from these countries successfully. This demonstrates the potential of a traceability database for cultured shrimp products from predominant production countries.

A hyperspectral imaging system was for the first time exploited to estimate the core colour of sausages stuffed in natural hog casings or in two hog casings treated with solutions containing surfactants and lactic acid in slush salt. Yellowness of sausages stuffed in natural hog casings (control group, 20.26 &plusmn, 4.81) was significantly higher than that of sausages stuffed in casings modified by submersion for 90 min in a solution containing 1:30 (w/w) soy lecithin:distilled water, 2.5% wt. soy oil, and 21 mL lactic acid per kg NaCl (17.66 &plusmn, 2.89) (p &lt, 0.05). When predicting the lightness and redness of the sausage core, a partial least squares regression model developed from spectra pre-treated with a second derivative showed calibration coefficients of determination (Rc2) of 0.73 and 0.76, respectively. Ten, ten, and seven wavelengths were selected as the important optimal wavelengths for lightness, redness, and yellowness, respectively. Those wavelengths provide meaningful information for developing a simple, cost-effective multispectral system to rapidly differentiate sausages based on their core colour. According to the canonical discriminant analysis, lightness possessed the highest discriminant power with which to differentiate sausages stuffed in different casings.

Incidents of fraud are emerging on the domestic Chinese honey market, especially where lower value rape honey is mixed with high-value acacia honey. In this study, we use a combination of 1H NMR spectroscopy combined with chemometric techniques to detect and quantify adulteration of acacia honey with cheaper rape honey. The honey 1H NMR spectroscopy was split into three regions i.e. the aliphatic (0.00–3.00 ppm), carbohydrate (3.00–6.00 ppm) and aromatic (6.00–9.50 ppm) regions, to investigate which region provided the best qualitative and quantitative indicators in terms of rape honey adulteration in acacia honey. Results showed that the highest prediction accuracy for rape honey addition is 89.7 % using canonical discriminant analysis (CDA), determined from compounds located in 0.00–6.00 ppm spectral range. Orthogonal projection to latent structures discriminant analysis (OPLS-DA) was used to further discriminate samples of pure acacia honey adulterated with different amounts of rape honey. A partial least squares model (PLS) established a linear fit between the actual and predicted adulterant concentration with an R2 value up to 0.9996. When the content of rape honey added was less than 100 g/kg, the purity of acacia honey could still be estimated using 1H NMR spectroscopy and chemometric methods.

Toxic pyrrolizidine alkaloids (PAs) and their N-oxides (PANOs) can be present in bee pollen depending on the plants visited by bees. A liquid chromatography-mass spectrometry (LC-MS/MS) method was developed and validated to monitor 17 PAs/PANOs in 44 bee pollens. The CIE-L∗a∗b∗ colour coordinates with the specular component either included or excluded were recorded in pellets and ground aliquots. Lightness (L∗) and yellowness (b∗) of ground bee pollen were significantly correlated to PAs/PANOs content. The L∗ and b∗ cut-offs sorted by a receiver operating characteristic analysis to predict PAs/PANOs presence showed a significant increase in the relative risk to detect amounts higher than 84 μg kg−1. Two supervised canonical discriminant analyses confirmed that pollen without PAs could be distinguished from those containing PAs/PANOs. The data suggest that instrumental colour coupled with supervised models could be used as a screening test for PAs/PANOs in bee pollen, before the confirmatory LC-MS/MS analysis., Food Science; Food Safety; Public Health; Toxicology; Analytical Chemistry; Bee pollen; Pyrrolizidine Alkaloids; LC-MS/MS; CIE-L∗a∗b∗ colour coordinates; receiver operating characteristic; canonical discriminant analysis.

The adulteration of minced mutton was studied by an electronic taste system with cross‐sensitive sensor array, providing a global liquid and taste perception to soluble flavour compounds in meat. The responses of taste sensors to adulterated mutton were collected, and analysed by multivariate data analysis methods. For discrimination of meat species and adulterated mutton with different content of pork/chicken, canonical discriminant analysis (CDA) and bayes discriminant analysis (BDA) and principle component analysis (PCA) were employed. The PCA and CDA results showed that meat of different species could be distinguished by E‐tongue responses. The CDA and BDA results showed effective classification results. For prediction of pork/chicken content in adulterated mutton, Multiple linear regression (MLR), Partial least square analysis (PLS) and Least Squares Support Vector Machines (LS‐SVM) were used, and the results were compared, finding that LS‐SVM was proved to be the most effective method for the prediction of pork/chicken content.

Capers and caper berries (Capparis spinosa L.) from different Sicilian localities and marked by Protected Geographic Indication (PGI) and Protected Designation of Origin (PDO), along with corresponding non Sicilian commercial samples, were subjected to several analytical and multivariate techniques in an attempt to predict the provenance of Capparis spinosa L. within the Sicilian region, and to discriminate Sicilian products from the non Sicilian counterpart. All samples were screened in terms of FA composition and elemental fingerprint. Then, principal component analysis (PCA) and canonical discriminant analysis (CDA) tested the capacity of both chemical classes to differentiate and classify capers and caper berries according to the geographical provenance. About 99% of total samples were discriminated by FAs, as they did not allow for a correct geographical classification of 7 samples of capers and related caper berries. Conversely, elements provided an excellent sample discrimination and prediction over the four locations (100%), revealing specific variables of classification (Mg, K, and Ca). For the first time, a reliable traceability system based on the elemental profiles of caper and caper berries was proposed. It could be useful for protecting the authenticity of Sicilian PGI and PDO capers and caper berries.

Although milk fat depression (MFD) has been observed and described since the beginning of the last century, all the molecular and biochemical mechanisms involved are still not completely understood. Some fatty acids (FA) originating during rumen biohydrogenation have been proposed as causative elements of MFD. However, contradictory results were obtained when studying the effect of single FA on MFD. An alternative could be the simultaneous evaluation of the effect of many FA using a multivariate approach. The aim of this study was to evaluate the relationship between individual milk FA of ruminal origin and MFD using canonical discriminant analysis, a multivariate technique able to distinguish 2 or more groups on the basis of a pool of variables. In a commercial dairy herd, a diet containing 26% starch on a DM basis induced an unintentional MFD syndrome in 14 cows out of 40. Milk yielded by these 14 animals showed a fat content lower than 50% of the ordinary value, whereas milk production and protein content were normal. The remaining 26 cows secreted typical milk fat content and therefore were considered the control group, even though they ate the same diet. The stepwise discriminant analysis selected 14 milk FA of ruminal origin most able to distinguish the 2 groups. This restricted pool of FA was used, as variables, in a run of the canonical discriminant analysis that was able to significantly discriminate between the 2 groups. Out of the 14 FA, 5 conjugated linoleic acid isomers (C18:2 trans-10,trans-12, C18:2 trans-8,trans-10, C18:2 trans-11,cis-13, C18:2 cis-9,cis-11, C18:2 cis-10,cis-12) and C15:0 iso were more related to the control group, whereas C18:2 trans-10,cis-12, C16:1 trans-6-7, C16:1 trans-9, C18:1 trans-6-8, C18:1 trans-9, C18:1 trans-10, C18:1 cis-11, and C18:3n-3 were positively associated with the MFD group, allowing a complete discrimination. On the basis of these results, we can conclude that (1) the shift of ruminal biohydrogenation from C18:1 trans-11 to C18:1 trans-10 seemed to be strongly associated with MFD; (2) at the same time, other C18:1 trans isomers showed a similar association; (3) on the contrary, conjugated linoleic acid isomers other than C18:2 trans-10,cis-12 seemed to be associated with a normal fat secretion. Results confirmed that MFD is the consequence of a combined effect of the outflow of many ruminal FA, which collectively affect mammary fat synthesis. Because the animals of the 2 groups were fed the same diet, these results suggested that factors other than diet are involved in the MFD syndrome. Feeding behavior (i.e., ability to select dietary ingredients in a total mixed ration), rumen environment and the composition of ruminal bacteria are additional factors able to modify the products of rumen biohydrogenation. Results of the present work confirmed that the multivariate approach can be a useful tool to evaluate a metabolic pathway that involves several parameters, providing interesting suggestions about the role of some FA involved in MFD. However, results about the MFD syndrome obtained in the present research require a deep molecular investigation to be confirmed.

Shrimp from three major exporting countries – India, Vietnam, and Thailand – were identified to country of origin by elemental profiling. Concentrations of 23 elements, including essential macro- and micro-nutrients and non-essential trace elements, in headless shell on shrimp (HLSO) samples were analyzed by ICP-AES. Elemental concentrations in shrimp showed high variation. Multivariate statistics including principal component analysis, stepwise discriminant analysis, Kernel method, and canonical discriminant analysis demonstrated the validity of elemental profiling in distinguishing aquacultured shrimp from different countries. However, it was not possible to reach definitive conclusions about which elements are the best descriptors for each region. The method was also tested as a means of differentiating shrimp produced to area of origin for two areas in Thailand, three provinces in Vietnam and India. All the multivariate statistical analysis demonstrated that the separation on a smaller scale was not as reliable as on the country scale.

The potential of diet-induced differences in the fatty acid profile of muscle to discriminate beef from different feeding systems and its potential use as an authentication tool was investigated. Three canonical discriminant models were built and validated using the fatty acid profile of beef from animals fed solely on pasture or cereal-based concentrates for 11 months or on various pasture/grass silage/concentrate combinations, including concentrates enriched with plant oils. Results indicated that models could successfully discriminate between grass-, partially grass- and concentrate-fed beef (accuracy = 99%) and between grass-fed beef and beef from animals supplemented with plant oils (accuracy = 96%). The approach also showed potential for distinguishing between beef from exclusively pasture-fed cattle and beef from cattle fed on pasture preceded by a period on ensiled grass (accuracy = 89%). Models were also applied to beef samples from 9 different countries. Of 97 international samples, including samples stated to be grass-fed, only 5% were incorrectly classified as Irish-grass-fed beef. These results suggested that the models captured traits in the fatty acid profile that are characteristic of Irish grass-fed beef and that this feature could be used for distinguishing Irish grass-fed beef from beef from other regions.

Phospholipid fatty acid (PLFA) profiles were studied for their potential to discriminate between freshwater-cultured and seawater-cultured salmonids. The effects of culture season and fish size on the PLFAs compositions in salmonids were investigated to assess their effects on the discrimination results. Totally 19 PLFAs of salmonids were determined via gas chromatography (GC). Among them, 12 PLFAs were significantly different between freshwater-cultured and seawater-cultured salmonids. But only four PLFAs in salmonids cultured in different seasons and two PLFAs in fish with different size were significantly different. The PLFAs, including 18:1n9, 24:1n9, 20:5n3, 20:3n6, ∑MUFA, and ∑n6PUFA, were selected via stepwise discriminant analysis (SDA) for further multivariate statistics analysis. Principal component analysis (PCA) and canonical discriminant analysis (CDA) visualized the differences of the PLFAs profiles in fish cultured in freshwater and seawater. The linear discrimination analysis (LDA) and random forest (RF) were used to create a discrimination model that can authenticate salmonids farmed in freshwater and seawater. An overall correct classification rate of 95.89%, a cross-validation rate of 93.15%, and a predictive classification rate of 100% were achieved by the LDA. The RF achieved an overall correct classification rate of 95.89%, a cross-validation rate of 95.89%, and a predictive classification rate of 100%. These results indicate that PLFAs profiles can be used to effectively distinguish samples cultured in freshwater and seawater while less affected by season and fish size. Therefore, PLFAs analysis is a promising method for the accurate identification of salmonids cultured in freshwater and seawater.

During ham processing the action of endogenous proteolytic and lipolytic enzymes leads to the development of volatile compounds (VOCs) responsible of typical aromas. Protected Designation of Origin (PDO) of Toscano ham requires at least 12 months of ripening but extended seasoning might improve flavor and economic value. This study aimed at assessing the evolution of color, fatty acids, and VOCs profile in subcutaneous fat, and, among VOCs, at identifying possible markers characterizing different seasoning length. For this purpose, a reduced pool of VOCs was selected by 3 multivariate statistical techniques (stepwise discriminant analysis, canonical discriminant analysis and discriminant analysis) to classify hams according to ripening (&lt, 12 months) or seasoning (&ge, 12 months) periods and also to seasoning length (S12, S14, S16, or S18 months). The main VOCs chemical families steadily increased along ripening. Aldehydes and hydrocarbons reached their peaks at S16, acids and ketones remained constant from R6 to S16, whereas esters started decreasing after 12 months of seasoning. Stepwise analysis selected 5 compounds able to discriminate between ripening and seasoning periods, with 1,1-diethoxyhexane and dodecanoic acid being the most powerful descriptors for ripening and seasoning period, respectively. Instead, 12 compounds were needed to correctly classify hams within seasoning. Among them, undecanoic acid methyl ester, formic acid ethyl ester, 2,4,4-trimethylhexane, and 6-methoxy-2-hexanone had a central role in differentiating the seasoning length.

The isotopic profile (δ(13) C, δ(15) N, δ(18) O, δ(2) H, δ(34) S) was used to characterise a wide selection of cocoa beans from different renowned production areas (Africa, Asia, Central and South America). The factors most influencing the isotopic signatures of cocoa beans were climate and altitude for δ(13) C and the isotopic composition of precipitation water for δ(18) O and δ(2) H, whereas δ(15) N and δ(34) S were primarily affected by geology and fertilisation practises. Multi-isotopic analysis was shown to be sufficiently effective in determining the geographical origin of cocoa beans, and combining it with Canonical Discriminant Analysis led to more than 80% of samples being correctly reclassified. Copyright © 2016 John Wiley & Sons, Ltd.

Anecdotal information suggests that some Hispanic consumers may consider US-made Hispanic cheeses as having a general lack of authenticity compared with those made in their countries of origin. To characterize the potential differences, samples of fresh, pasta filata, and aged Hispanic cheeses were acquired from both the United States (total n=39) and countries of origin (total n=30) purchased from Mexico, Central America (Costa Rica and El Salvador), and the Caribbean (Puerto Rico). The proximate composition, microbial counts, melt profile, and sensory characteristics were evaluated and compared in country-of-origin cheeses and the US-made counterparts. The presence of Listeria spp. was confirmed for 1 Mexican aged cheese sample and 6 cheese samples from Central America (3 fresh, 2 pasta filata, and 1 aged). The chemical composition, melt profile, and sensory characteristics of fresh and pasta filata US Hispanic cheeses were not significantly different from their Mexican counterparts. Likewise, the chemical composition and melt profile of US aged Hispanic cheeses was not significantly different from the aged Mexican cheeses, but sensory characteristics varied among all aged cheeses. These results demonstrate the similarities and differences among US fresh, pasta filata, and aged Hispanic cheeses relative to their counterparts made in the countries of origin.

A study is presented of the real possibilities of glucosinolate content and chemometrics (canonical discriminant analysis) to differentiate bee pollen samples from four different apiaries (Fuentelahiguera, Monte, Pistacho, Tio Natalio) located in the same geographical area. Fifteen intact-glucosinolates were quantified by means of ultra-performance liquid chromatography coupled to a quadrupole time-of-flight mass detector in forty-nine bee pollen samples. Glucosinolate residues were detected in most of the samples, and these differed in number and concentration. It was possible to directly differentiate one of the apiaries (Fuentelahiguera) from the other three (Monte, Pistacho y Tio Natalio) by comparing glucosinolate content. These three apiaries were differentiated by means of the first two canonical variables obtained from a canonical discriminant analysis. Following this analysis, more than 88% of the samples could be assigned correctly to the Pistacho and Monte apiaries, and 100% to the Tio Natalio apiary.

The fatty acid (FA) profiles of lamb muscle and adipose tissue originating from ten different grazing areas in North Macedonia and an assessment of whether the meat origin could be distinguished on the basis of FA profile are presented. Muscle and adipose tissues of three-month-old male lambs (crossbreds of autochthonous Ovcepolian and Wurttemberg) reared on spring pastures were studied. Statistically significant differences in the individual FA contents, FA groups and FA ratios were observed between lamb tissues from the different geographic regions. Canonical discriminant analysis showed there was a significant linear divergence between tissues from almost all examined regions. The greatest weight in the differentiation of the different areas showed fatty acid ratios, C18:3n3, C18:1n9c and C20:5n3 for the muscle tissue, as well as C18:1n9c, C18:2n6c, C16:1 and C17:1 for the adipose tissue. This showed that the FA composition of muscle and adipose tissue obtained from lambs reared on pasture could serve as a useful indicator to aid differentiation of its geographic origin but it should be confirmed through further replicated studies.

Salmonid samples collected from two sites in different aquaculture areas (Yantai and Liujiaxia, China) in four seasons were subjected to multi-element analysis. The amounts of 18 elements in fish were measured by inductively coupled plasma atomic emission spectrometry (ICP-AES). The results showed that concentrations of elements in fish from Yantai were stable with seasonal alternation. However, the element concentrations and compositions of fish obtained from Liujiaxia were vulnerable to seasonal change. Principal component analysis (PCA) and canonical discriminant analysis (CDA) were used to visualize the regional and seasonal distribution of samples, and it was determined that CDA was more distinct than PCA. To determine if seasonal effects would influence the discrimination of the geographical origin of salmonid, multivariate statistics including linear discriminant analysis (LDA), k-nearest neighbor (KNN), and partial least squares discriminant analysis (PLS-DA) were used to discriminate fish samples from the two different areas. The results showed that all discrimination techniques could effectively distinguish samples while remaining unaffected by seasonal effects.

The administration of anabolic agents in farm animals to improve meat production has been prohibited in EU, due to the potential risks to human health. Meat quality was investigated to detect the effects of illegal administration of dexamethasone or prednisolone or 17β-estradiol on Charolais bulls. Three groups of 6 bulls were treated and 12 bulls were the control. Meat quality parameters were measured on live animals, carcasses and on samples of Longissimus thoracis and multivariate statistical data analysis was applied. In Charolais bulls, these parameters were affected by growth promoter administration and the multivariate canonical discriminant analysis was able to distinguish between treated and untreated animals mainly due to three electronic nose’s parameters, 24 h carcass temperature and drip loss. Therefore, meat quality control and the multivariate analysis could be useful as a first screening to address targeted controls on farms suspected of illicit use of growth promoters.

In the last years, element content has been used as geographical tracer to determine the provenance of food. In the present work the content of 19 minerals (K, Ca, Mg, Na, Fe, Zn, B, Cu, Al, Mn, Ni, Cr, Pb, Co, As, Se, Cd, Sb and V) and 13 lanthanide elements (La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm and Lu) in Protect Geographical Indication (PGI) Interdonato lemon Messina (Italy) and non-PGI Turkish lemon pulps was evaluated by ICP-MS. The collected data were used to discriminate geographical origin of lemon samples (PGI or non-PGI) through statistical analyses. The discrimination between Italian and Turkish lemons was achieved by Principal Component Analysis (PCA) and also by Canonical Discriminant Analysis (CDA). The results indicate that the 100% of total samples were correctly classified. The present study suggests that it is possible to relate lemons to their geographical origins, representing a necessary condition for traceability of these peculiar products. Unfortunately, Italian and Turkish lemon pulps cannot be considered “rich in” or “source of” analyzed minerals. However, these fruits can be considered safe for their low content in toxic elements.

UNLABELLED The profiles of phenolic acids and flavan-3-ols for the selected Chinese red wines and the potential of using phenolic acids and flavan-3-ols to differentiate the geographic origin and grape variety of wines from China are investigated in this study. Significant differences and markers could be found according to the geographical origin and grape variety. Through a canonical discriminant analysis a good differentiation was developed according to the geographic origin or grape variety, and the accuracy of the discriminant model was 88.9% and 100%, respectively. According to the phenolic acid and flavan-3-ols profiles of the wine samples and good differentiation in the region and the variety discriminant analysis, minimal fraudulent claims were noted for the Chinese red wines investigated. PRACTICAL APPLICATION This study provides some help for the protection of geographical origin and monovarietal wine claims.

Espresso coffee beverages prepared from pure origin roasted ground coffees from the major world growing regions (Brazil, Ethiopia, Colombia, India, Mexico, Honduras, Guatemala, Papua New Guinea, Kenya, Cuba, Timor, Mussulo and China) were characterized and compared in terms of their mineral content. Regular consumption of one cup of espresso contributes to a daily mineral intake varying from 0.002% (sodium; Central America) to 8.73% (potassium; Asia). The mineral profiles of the espresso beverages revealed significant inter- and intra-continental differences. South American pure origin coffees are on average richer in the analyzed elements except for calcium, while samples from Central America have generally lower mineral amounts (except for manganese). Manganese displayed significant differences (p < 0.05) among the countries of each characterized continent. Intercontinental and inter-country discrimination between the major world coffee producers were achieved by applying canonical discriminant analysis. Manganese and calcium were found to be the best chemical descriptors for origin.

Elemental analysis (Al, Ca, Cr, Cu, Fe, K, Mg, Mn, Na, P, S, Se, and Zn) of fillets of channel catfish Ictalurus punctatus and hybrid catfish (♀ I. punctatus × ♂ blue catfish Ictalurus furcatus) from fertilized ponds and of hybrid catfish that received feed was performed using an inductively coupled plasma atomic emission spectrometry (ICP-AES). Fillets of samples of hybrid and channel catfish in fertilized ponds did not differ in composition for most of the elements analyzed. Multivariate statistical methods including principal component analysis (PCA), canonical discriminant analysis (CDA) and k-nearest-neighbor analysis were used to separate Ictalurid catfish to the two culture methods based on chemical composition. The three methods each demonstrated that fillets from fish that received feed could be differentiated from fillets of fish cultured in fertilized ponds. Fish from fertilized ponds rely on natural food organisms in the same manner as wild-caught fish. This study demonstrated the potential of chemical profiling combined with multivariate statistical methods for verifying method of production – aquaculture versus wild-caught.

Different varieties of rice grown in various countries around the world (Australia, China P.R., France, India, Italy, Japan, Korea, Malaysia, Myanmar, Pakistan, Spain, Taiwan, Thailand, USA and Vietnam) were analysed using isotope ratio mass spectrometry and inductively coupled plasma mass spectrometry. The stable isotope ratios of carbon, nitrogen, oxygen and hydrogen and the multi-elemental compositions were assessed as variables for discrimination of the geographical origins of the rice samples. The data were processed by canonical discriminant analysis (CDA) enabling classification according to geographical origin. Fifteen key variables were identified by CDA as providing the maximum discrimination between the rice samples across different rice types and categorised on the basis of broad geographical areas (Asia, Australia, Europe, India and Pakistan, North America and Southeast Asia), enabling 90.7% correct classification for the model generated. This feasibility study demonstrated that the methodolo...

The aim of this study was to differentiate between wines of three diverse wine regions, namely the districts Paarl, Stellenbosch and Swartland, in the Western Cape on the basis of chemical parameters, namely pH, volatile acid, total acids, malic acid, lactic acid, glucose, fructose, ethanol and glycerol. A total of 140 young red wines (Pinotage, Cabernet Sauvignon and Merlot) obtained from the 2003 National Young Wine Show were analysed. Classification methods such as stepwise discriminant analysis, canonical discriminant analysis and linear discriminant analysis were applied to the data sets. A total classification of 34.04 % for Cabernet Sauvignon, 42.55 % for Pinotage and 31.25 % for Shiraz was determined. It was concluded that, under the conditions of this study, separation of the wines according to geographical regions was unsuccessful using the selected parameters.

1. The skulls and postcranial skeletons of the red jungle fowl (Gallus gallus) were compared osteometrically between the populations from North and South Vietnam, North and Central Laos and Southeast Bangladesh. The populations include the three subspecies of G. g. spadiceus, G. g. gallus and G. g. murghi and were sampled to reveal the geographical morphological variations among populations in G. gallus. 2. The morphometric characteristics of subspecies murghi could be clearly distinguished from those of the other subspecies using a canonical discriminant analysis. However, the size and shape of the skull of the gallus population from South Vietnam were not statistically different from that of the subspecies spadiceus from North Laos. The canonical discriminant scores also clearly indicated that there were morphological similarities in the skulls of the populations from North Laos and South Vietnam. 3. From the results, therefore, it is concluded that red jungle fowls do not exhibit high levels of osteometric variation between geographical localities at least within the Indochinese Peninsula. 4. This contrasts with previous studies which have described these subspecies as having various external morphological differences and have argued that zoogeographical barriers exist between the north and south areas of the Indochinese Peninsula.

Frutos de 86 ecotipos de ameixa mexicana foram colhidos dos estados de Guerrero e Morelos durante a estacao seca. Destes frutos, 22 foram ecotipos selvagens e 64 variedades cultivadas. Entre as variaveis medidas, aquelas com os maiores coeficientes de variacao foram cor, sabor e massa (45), destacando-se a presenca de consideravel variabilidade intra-especies. A analise de clusters separou os 86 acessos em 5 grupos, principalmente com base na cor, sabor, comprimento e massa. Os membros dos tres primeiros grupos apresentaram epicarpos vermelhos (Grupo I), amarela (Grupo II) ou purpura (Grupo III) e maiores valores de massa (12,2-16 g), comprimento (29,6-33,9 mm), rendimento de polpa (68,8 -71,9), TSS (11,16-11,52 Brix) e indice de aroma (14,5-18,3), tornando-os adequados para uso horticola e consumo fresco. Os ecotipos selvagens se agruparam nos restantes dois grupos e consistiram em pequenas drupas vermelhas (23,2-27,7 mm, 5,5-8,2 g) de diferentes tonalidades. A cor vermelha cereja dos frutos do Grupo IV sugere possiveis propriedades antioxidantes devido a presenca de pigmentos polifenolicos que poderiam ser de interesse para as industrias farmaceutica e cosmetica. Por fim, os frutos do Grupo V, sendo os mais acidos (pH 2,7- 2,1 de acidez), podendo ser mais adequados para a preparacao de conservas e molhos

The lipid fraction of Asiago Protected Designation of Origin (PDO) cheese was analyzed to identify specific biomarkers of its main production systems through a canonical discriminant analysis. The three main production systems of the cheese were considered. Two were located in the upland (UL): pasture-based (P-UL) vs hay-based total mixed rations (H-UL). The third was located in the lowland (LL) and processed milk from cows fed maize silage-based rations (maize silage lowland: MS-LL). The discriminant analysis selected nine fatty acids and vitamin A as lipid biomarkers useful to separate the three production systems. High contents of conjugated linoleic acids, anteiso-C15:0, and vitamin A were discriminant factors for P-UL cheese. The separation between H-UL and MS-LL cheese was less marked with the former having the higher content of conjugated linoleic acids and some polyunsaturated n-6 fatty acids and with the latter being identified by cyclopropane fatty acid and C9:0.

This study aimed to investigate the triacylglycerol (TAG) composition of Asian sesame oils and to differentiate the geographic origins of the oils using the analytical data in combination with canonical discriminant analysis (CDA). The TAG compositions of 51 Korean, 19 Chinese, and 14 Indian sesame oil samples were analyzed by liquid chromatography-evaporative light scattering detector (LC-ELSD) and GC-FID, separately. The TAG content values determined by LC-ELSD showed strong positive linear correlation to those measured by GC-FID. The CDA results showed that good discrimination (hit ratio, ∼92%) between the oil samples was achieved by applying the respective canonical variables generated from the LC-ELSD and GC-FID data. The CDA using GC-FID data accurately characterized the origins of all the four additional commercial oil samples (two Korean and two Chinese products), whereas the CDA using LC-ELSD data correctly characterized the origins of only two samples. This study demonstrated that both LC-ELSD and GC-FID analyses of TAG compositions were useful tools for characterizing the geographic origin of the oils and the separation of the Korean sesame oils from the other two groups of oils using the GC-FID data was more efficient than that using the LC-ELSD data. Practical applications: The simplicity and universality of GC-FID and LC-ELSD make the analysis of TAG composition using these chromatographic systems a practical and feasible procedure for tracing the geographic origin of the sesame oils. The combination of TAG analysis with methods to analyze different constituents in sesame oils would enhance the accuracy of the oil traceability. Score plots of two canonical variables derived from CDA using liquid chromatography-evaporative light scattering detector (LC-ELSD) data and GC-FID data for the Korean, Chinese, and Indian roasted sesame oils.

a b s t r a c t A nondestructive method based on visible and near infrared spectroscopy, was investigated for deter- mining the maturity of green tomatoes at harvest. The interactance spectrum of tomato fruit at the firm green stage were measured in less than 5 s with a handheld spectrometer (400-1000 nm) utilizing an optical design with a 0◦ angle between the incident illumination and the detection view. Results showed that the predominant change in the interactance spectra in green tomatoes useful for predicting chang- ing maturity levels occurred in the 600-750 nm portion of the 400-1000 nm region, typically associated with chlorophyll. Variety-specific Bayesian classification models and a joint variety 'global' classification model were developed to predict tomato maturity after 7 d of storage in a 20 ◦C ripening room using canonical discriminant analysis techniques applied to the interactance spectra from 600 nm to 750 nm. Variety specific models correctly identified 75-85% of immature tomatoes and 82-86% of mature green tomatoes in internal cross-validation, however external validation performance decreased when applied to predict maturities in a cultivar external to the training set. False positive rates of these models in identifying immature and mature green tomatoes varied from 3% to 40% and 0% to 31% respectively. A 'global' model, trained on two cultivars showed more stability and correctly predicted 71% of immature and 85% of mature green tomatoes, with false positive error rates of 13% and 22%, respectively, in internal cross-validation of both varieties. This handheld system showed good potential as a rapid, nondestructive technique to aid tomato production managers in the identification of immature green tomatoes at har- vest and could be a valuable tool in delivering more flavorful fruit to consumers by reducing the amount of immature fruit harvested by workers. © 2013 Published by Elsevier B.V.

In order to verify the authenticity of organic products, multivariate data (δ15N values, key-metabolites and mineral content) were used to classify organic and conventional peppers and lettuce by means of canonical discriminant analysis. Different origins and types of fertilisers were considered. This is the first paper to look into the influence of biosolarisation, a soil disinfection technique widely used in organic cultivation, on the differentiation between organic and conventional crops. The use of pre-harvest parameters (plant and soil δ15N and plant weight) allowed 78.8% and 65.4% of pepper and lettuce samples, respectively, to be correctly classified. In pepper, this percentage was reduced (69.2%) when only post-harvest sampling parameters were used to build the classification model. In lettuce, the use of biomarkers significantly improved the model, allowing 90.4% of cases to be correctly classified. This methodology showed a high potential for use as an authenticity control tool for organic cultivation.

Bitter pit (BP) is a physiological disorder in apple commonly associated with high Mg/Ca ratio in fruit tissue. This work was carried out to identify the best fruit tissue sampling method for Ca and Mg assessment to discriminate ‘Fuji’ apples regarding the occurrence of BP. After six months under controlled atmosphere storage (2.0 kPa O2 + 0.5 kPa CO2, at 0.5 °C/90–95% RH), fruit without or with BP were submitted to Ca and Mg analysis (mg kg−1 of fresh weight) in the total fraction (TF) of peel + flesh, and in the TF and soluble fraction (SF) of flesh or peel tissues. For the peel + flesh, a wedge-shaped segment was cut longitudinally from the fruit (with 1 cm wide at the equatorial region), discarding the core tissue. For individual sampling of the peel (thickness of 0–2 mm) or flesh (thickness of 2–8 mm), the fruit were cut along the equatorial region, so that only the distal end was used. In this portion of the fruit, Ca and Mg contents were higher in the peel than in the flesh in the TF, and lower in the peel than in the flesh in the SF, while the Mg/Ca ratios in TF and SF fractions were higher in the flesh than in the peel, regardless the occurrence of BP. Calcium and Mg contents in the SF relative to the TF were very low in both peel and flesh tissues. Calcium in the SF represented 0.36% and 2.79% of its content in the TF for peel and flesh, respectively. Magnesium content in the SF represented 0.70% and 3.74% of that in the TF for peel and flesh, respectively. Fruit without BP showed higher Ca content and lower Mg/Ca ratio in the TF of peel + flesh, peel or flesh, and in the SF of peel or flesh, compared to fruit with BP. Fruit without BP also showed a higher percentage of Ca in the SF in relation to TF in the peel (0.42%) compared to fruit with BP (0.31%). Fruit without BP also showed lower Mg content in TF of peel + flesh and peel, and in the SF of the flesh. There was no difference between fruit with and without BP only for Mg contents in TF of the flesh and SF of the peel. The canonical discriminant analysis showed that the Mg/Ca ratio in the SF of the peel tissue at the distal end of the fruit provides the best method to discriminate fruit without and with BP, which can be potentially used as a tool to predict BP development in ‘Fuji’ apples.

Volatile compounds in the peel of calamansi (Citrus microcarpa) from Malaysia, the Philippines and Vietnam were extracted with dichloromethane and hexane, and then analysed by gas chromatography–mass spectroscopy/flame ionisation detector. Seventy-nine compounds representing >98% of the volatiles were identified. Across the three geographical sources, a relatively small proportion of potent oxygenated compounds was significantly different, exemplified by the highest amount of methyl N-methylanthranilate in Malaysian calamansi peel. Principal component analysis and canonical discriminant analysis were applied to interpret the complex volatile compounds in the calamansi peel extracts, and to verify the discrimination among the different origins. In addition, four common hydroxycinnamic acids (caffeic, p-coumaric, ferulic and sinapic acids) were determined in the methanolic extracts of calamansi peel using ultra-fast liquid chromatography coupled to photodiode array detector. The Philippines calamansi peel contained the highest amount of total phenolic acids. In addition, p-Coumaric acid was the dominant free phenolic acids, whereas ferulic acid was the main bound phenolic acid.

A discrimination technique for domestic and imported Scutellaria baicalensis was developed using an energy dispersive X-ray fluorescence spectrometer (ED-XRF). Mineral content ratios, of a total of 43 species, including P, S, Cl, K, Ca, Mn, Fe, Cu, and Zn, were measured among 204 samples. Macro element content ratios and trace element content ratios were determined using the standardless fundamental parameters (SLFP) analysis. Inorganic element ratios of P, S, K, Ca, Cl, Mn, and Fe were significantly different between domestic and imported samples. The result from the canonical discriminant analysis showed that the accuracy of geographical origin discrimination was 95.15%; the correlation coefficient was 0.888. It was concluded that this technique could be used as a useful method in discriminating the geographical origins between domestic and imported Scutellaria baicalensis.

Summary The anthocyanin composition of Tannat red wines obtained with four winemaking techniques was evaluated in different years. The wines were elaborated with traditional maceration (TM), cold pre-fermentative maceration, delayed extraction of anthocyanins and extended maceration. Two vinifications were carried out for each technique, employing 70 kg of grapes in each one. The anthocyanin composition of wines was analysed by high-performance liquid chromatography–diode array detection. The results confirm the fundamental role of the grape composition on the anthocyanin composition of the wines. The winemaking had an important effect on the total concentration of anthocyanins and also may modify the anthocyanic profile of wines. The wines obtained with TM had higher or similar content of anthocyanins than those elaborated with the other techniques. Canonical discriminant analysis of data showed that vintage has a greater effect on anthocyanin concentration than winemaking techniques.

Mg, Ca, Zn, Cu, K and Na were determined in some species of Slovakian fruit distillates by atomic absorption spectrometry with the aim to differentiate the spirit drinks according to geographical origin. Potassium, sodium and copper were found as markers with the highest concentrations and variability in the distillates, namely in the apricot and grape brandy. Using the multivariate statistics of principal component and canonical discriminant analysis enabled relative effective differentiation of samples according to their regional origin. Prediction ability of the model resulted in more than 80% of correctly classified samples of the fruit distillates into the relevant Slovakian regions.doi: 10.5219/164

Introduction Mentha pulegium L. (pennyroyal) is one of the four most commercially important Mentha species, even it is not a cultivated plant. It can be abundantly located in the Iberian Peninsula and North African countries. In Greece it grows in the wild and it is scattered all over the country. Pennyroyal is best known for its essential oil, with Spain and Morocco being the largest producers in the world. Mid-infrared spectroscopy has been applied to determine the origin of various samples. Objectives In this work Fourier transform infrared spectroscopy (FT-IR) combined with canonical discriminant analysis has been applied to distinguish 70 Greek pennyroyal samples according to their collection areas. Material and methods Pennyroyal nonpolar organic extracts were prepared using ultrasound-assisted solvent extraction. The spectra of the extracts were recorded in the range of 4000–400 cm−1 and the best discrimination was achieved in the spectral region 1720–1650 cm−1. Results Spectral features for the discrimination of pennyroyal samples among the different collection areas occur primarily in the carbonyl region and are correlated with the main volatile constituents of the extracts (menthone, isomenthone, pulegone, piperitone). All areas were easily differentiated by canonical discriminant analysis. The percentages of correct classification and validation were 94.3 and 90.0%, respectively. Conclusion The combination of FT-IR spectroscopy and multivariate analysis provides a rapid and ambient method to discriminate pennyroyal samples in terms of geographical origin. Copyright © 2011 John Wiley & Sons, Ltd.

Effects of the Korean carcass-grading system on carcass traits and meat quality parameters for the longissimus dorsi (LD) and semimembranosus (SM) muscles were investigated for Hanwoo steer beef. High quality grade beef cuts had significantly higher intramuscular fat content than those of low quality grade beef cuts. To establish a guaranteed system of eating quality for Hanwoo steer beef, a palatability prediction model was developed. The model development approach used canonical discriminant analysis and multivariate regression to assess effects of muscle type, cooking method, and aging treatment on Hanwoo beef quality grade. The palatability prediction model can provide Korean consumers with detailed information about expected eating quality for an individual cut, as well as provide critical information to the industry for maintaining high-quality beef production. The model is a significant advance in grading compared with the traditional carcass grading system, which applies a single quality grade to an entire carcass.

The potential of using multi-spectral vision technology for quality control in a continuous frying process was investigated. Canonical discriminant analysis of the multi-spectral images of samples of fried minced meat and diced turkey could clearly visualise the effect of different heat treatments. The vision technology can also detect even slight increases in the agglutination of the fried minced meat during the process. This agglutination is undesirable, but very difficult to measure on-line. The results indicate that multi-spectral vision technology may partially or totally substitute visual inspection by an operator as a means of assessing product quality during processing.

The objective of the research was to compare the taste attributes of different commercial soya drinks. Furthermore, the task was to determine the effect of different ingredients and processing technologies on the taste attributes of the product. Based on the results of electronic tongue measurements the instrument is able to determine the effect of the applied technology and to distinguish soya juice samples according to sensory preferences. Canonical discriminant analysis showed that the groups of two measurements of the same products were overlapping. Therefore, the electronic tongue measurements are supposed to be of acceptable repeatability. The canonical discriminant analysis showed that the taste attributes of soya juice made of hulled soybeans was beneficial for the taste attributes relative to that of the juice made of not-hulled soybeans. Three main groups could be observed from the analyzed six commercial soya drink samples based on canonical discriminant analysis. There is a group of top market brands having definite taste improver additives and another one containing three products having low amount of additives. However, the group of samples made of soybean and rice is located between the above-mentioned two groups in the discriminant score plot.

Akara (fried cowpea paste) and moinmoin (steamed cowpea paste) are foods traditionally prepared from dehulled and wet-milled seeds. These products were prepared using flour produced from five varieties of cowpea as starting ingredient. The sensory properties of products from flour were compared with those from fresh paste (control) using the Difference from Control Test. Trained panelists (n = 9) rated samples on an 8-level scale, where 0 = no difference and 7 = extremely large difference at triplicate sessions on separate days for each product. Results showed significant differences (α ≤ 0.05) in mean scores for attributes of products among flour of different cowpea varieties. Mean scores ranged between 0.30 (no difference) to 3.1 (moderate difference) for all attributes of akara, while for moinmoin, a range of 1.0 (very slight difference) to 3.4 (moderate difference) were obtained. Canonical discriminant analysis showed that the most discriminating sensory attributes of akara were crumb sponginess, lightness and crust softness, while those of moinmoin were mouthfeel and moistness. Based on the most discriminating attributes, the varieties California Blackeye 5 and Mississippi silver gave products that were the least different from the control. PRACTICAL APPLICATIONS Akara and moinmoin are popular West African foods prepared from cowpea paste, which require laborious and time-consuming operations to process from seeds. The availability of convenient flour that can be simply rehydrated would greatly simplify the preparation of these foods. However, the acceptance of cowpea flour in the marketplace as an alternative starting ingredient depends on the retention of requisite functionality that would give the desired sensory quality in the products. The sensory attributes of akara and moinmoin identified in the study are those that must not be compromised by the conditions of processing flour from seeds. This knowledge would be useful in the development of an appropriate flour-processing method.