Ana C. Pereira

Aroma ageing trends in GC/MS profiles of liqueur wines.

Madeira wine has been studied with the main goal of acquiring a better understanding about the evolution of its properties over time. For that purpose, flexible and reliable data analysis tools were employed to characterize wines at different ageing stages, using flavour chromatography measurements. In this paper we present the results from such a study, where the main differences in the aroma profiles and their development in different types of aged Madeira wines are analyzed and evaluated according to their discriminating power. An exploratory multivariate data analysis was conducted using two different tools, namely biplots and contributions plots obtained through principal component analysis (PCA). In order to take advantage of the maximum amount of information provided by the chromatography data sets, a new approach that incorporates samples variability in the analysis of the statistical significance of contributions estimates, was developed and tested. In this way, it was possible to analyze which volatile compounds have statistically significant and/or similar contributions regarding the observed separation of wine samples from different groups. Furthermore, since several chemical compounds are expected to change together as a result of the ageing-related chemical reactions, they were clustered according to a similarity criterion relative to their importance in the trends observed in the scores space. Results obtained provide a sound basis for the differentiation and characterization of the ageing process followed by Madeira wines.

Rapid and sensitive methodology for determination of ethyl carbamate in fortified wines using microextraction by packed sorbent and gas chromatography with mass spectrometric detection

This work presents a new methodology to quantify ethyl carbamate (EC) in fortified wines. The presented approach combines the microextraction by packed sorbent (MEPS), using a hand-held automated analytical syringe, with one-dimensional gas chromatography coupled with mass spectrometry detection (GC-MS). The performance of different MEPS sorbent materials was tested, namely SIL, C2, C8, C18, and M1. Also, several extraction solvents and the matrix effect were evaluated. Experimental data showed that C8 and dichloromethane were the best sorbent/solvent pair to extract EC. Concerning solvent and sample volumes optimization used in MEPS extraction an experimental design (DoE) was carried out. The best extraction yield was achieved passing 300 μL of sample and 100 μL of dichloromethane. The method validation was performed using a matrix-matched calibration using both sweet and dry fortified wines, to minimize the matrix effect. The proposed methodology presented good linearity (R(2)=0.9999) and high sensitivity, with quite low limits of detection (LOD) and quantification (LOQ), 1.5 μg L(-1) and 4.5 μg L(-1), respectively. The recoveries varied between 97% and 106%, while the method precision (repeatability and reproducibility) was lower than 7%. The applicability of the methodology was confirmed through the analysis of 16 fortified wines, with values ranging between 7.3 and 206 μg L(-1). All chromatograms showed good peak resolution, confirming its selectivity. The developed MEPS/GC-MS methodology arises as an important tool to quantify EC in fortified wines, combining efficiency and effectiveness, with simpler, faster and affordable analytical procedures that provide great sensitivity without using sophisticated and expensive equipment.

Optimal design of experiments applied to headspace solid phase microextraction for the quantification of vicinal diketones in beer through gas chromatography-mass spectrometric detection

Vicinal diketones, namely diacetyl (DC) and pentanedione (PN), are compounds naturally found in beer that play a key role in the definition of its aroma. In lager beer, they are responsible for off-flavors (buttery flavor) and therefore their presence and quantification is of paramount importance to beer producers. Aiming at developing an accurate quantitative monitoring scheme to follow these off-flavor compounds during beer production and in the final product, the head space solid-phase microextraction (HS-SPME) analytical procedure was tuned through experiments planned in an optimal way and the final settings were fully validated. Optimal design of experiments (O-DOE) is a computational, statistically-oriented approach for designing experiences that are most informative according to a well-defined criterion. This methodology was applied for HS-SPME optimization, leading to the following optimal extraction conditions for the quantification of VDK: use a CAR/PDMS fiber, 5 ml of samples in 20 ml vial, 5 min of pre-incubation time followed by 25 min of extraction at 30 °C, with agitation. The validation of the final analytical methodology was performed using a matrix-matched calibration, in order to minimize matrix effects. The following key features were obtained: linearity (R(2) > 0.999, both for diacetyl and 2,3-pentanedione), high sensitivity (LOD of 0.92 μg L(-1) and 2.80 μg L(-1), and LOQ of 3.30 μg L(-1) and 10.01 μg L(-1), for diacetyl and 2,3-pentanedione, respectively), recoveries of approximately 100% and suitable precision (repeatability and reproducibility lower than 3% and 7.5%, respectively). The applicability of the methodology was fully confirmed through an independent analysis of several beer samples, with analyte concentrations ranging from 4 to 200 g L(-1).

Advanced predictive methods for wine age prediction: Part II – A comparison study of multiblock regression approaches

In this article, we extend the scope of the first paper of the sequel, which was dedicated to the analysis of advanced single-block regression methods (Rendall et al., 2016) [1], to the class of multiblock regression approaches. The datasets contemplated for developing the multiblock approaches are the same as in the former publication: volatile, polyphenols, organic acids composition and the UV-Vis spectra. The context is still the prediction of the ageing time of one of finest Portuguese fortified wines, the Madeira Wine, but now the data collected from the different analytical sources is explored simultaneously, in a more structured and informative way, through multiblock methodologies. The goal of this paper is to provide a critical assessment of a rich variety of multiblock regression methods, namely: Concatenated PLS, Multiblock PLS (MBPLS), Hierarchical PLS (HPLS), Network-Induced Supervised Learning (NI-SL) and Sequential Orthogonalised Partial Least Squares (SO-PLS). A comparison of block scaling methods was also undertaken for the Concatenated PLS algorithm, and new block scaling methods were proposed that led to better prediction performances. This study explores and reveals the potential advantages of applying multiblock methods for fusing datasets from different sources, from both the predictive and interpretability perspectives.

Amino Acids and Biogenic Amines Evolution during the Estufagem of Fortified Wines

The current study was focused on the impact of accelerated ageing (heating step) on the amino acid and biogenic amine profiles of fortified wines. In this sense, three Madeira wines from two commonly used grape varieties (one red and the other white) were analysed during the heating, at standard (45°C, 3 months) and overheating (70°C, 1 month) conditions, following a precolumn derivatization procedure using iodoacetic acid, o-phthaldialdehyde, and 2-mercaptoethanol, carried out in the injection loop prior to RP-HPLC-FLD detection. Eighteen amino acids were identified, with arginine being the most abundant. An important decrease of the amino acid levels was detected during the standard heating (up to 30%), enhanced up to 61% by the temperature increase. Cysteine, histidine, and asparagine revealed the greatest decreases at 45°C. Conversely, some amino acids, such as asparagine, slightly increased. Four biogenic amines were identified but always in trace amounts. Finally, it was observed that the accelerated ageing did not favour the biogenic amine development. The results also indicate that the heating process promotes the amino acid transformation into new ageing products.

Nutritional and Phytochemical Composition of Vaccinium padifolium Sm Wild Berries and Radical Scavenging Activity

Blueberries have a well-deserved reputation as a potential functional food, supported by studies which have identified and quantified various nutrients and bioactive phytochemicals with known benefits for human diet and health. Wild blueberries have attracted particular attention due to the levels and concentrations of those phytonutrients. This study aims to evaluate for the first time the chemical composition of Madeira Islands endemic Vaccinium padifolium Sm wild berry. Results show that this fruit contains high values of total soluble phenolic content (around 4 g GAE kg-1 FW), as well as significant values of total monomeric anthocyanin content (around 3 g eq. cyanidin kg-1 FW) and DPPH scavenging activity (around 86.72%). Additionally, results reveal that this fruit has water content of about 88% as well as low sugar content (17.98 and 29.73 g kg-1 for glucose and fructose, respectively). Results also confirm that this wild blueberry is a good source of dietary fiber, fat and minerals. The high level of terpenoid compounds stands out in the aroma profile analysis. PRACTICAL APPLICATION This study is in line with the efforts of the scientific community to identify new sources of phytonutrients that are beneficial to human health, characterizing the wild Madeira blueberry in terms of phytonutrients that suggest there may be health benefits associated with its consumption. The findings of this research are very important for both the commercial and agricultural sectors that produce this fruit, and for consumers who seek phytonutrient-rich foods.

Multivariate Statistical Monitoring of Wine Ageing Processes

Abstract The flavor pattern is a key quality feature in the wine industry. Being the result of a complex interplay of different classes of volatile compounds, it presents an important evolution during the final and longer phase of the wine production process: the ageing period. In this paper, we present a data analysis framework for supporting the proper monitoring of the quality of wine products, during the ageing process, focusing on their flavor characteristics. We focus our analysis on a high quality Portuguese wine, the fortified Madeira wine, for which samples were collected over an extended time period, and analyzed in terms of their flavor composition using GC-MS. Then, several classification methodologies for age prediction were developed and evaluated, after a preliminary feature extraction stage using partial least squares for discriminant analysis (PLS-DA). Our analysis shows that it is indeed possible to identify the relevant ageing patterns and trends in a lower dimensional subspace, and to achieve very interesting classification performances, despite the natural variability present in wine products. The proposed framework also offers the potential to be applied in identity assurance and fraud detection tasks.

An extended comparison study of large scale datadriven prediction methods based on variable selection, latent variables, penalized regression and machine learning

Abstract Regression methods are pervasive in most data-driven predictive studies performed with industrial and laboratorial data. They provide the means to obtain reliable estimates of output variables (related to product quality or other properties) based on a set predictor variables which are usually easier to measure and less expensive. In this paper, a large scale comparison framework is developed in order to assess the performance of a rich variety of regression methods, compare them and provide guidelines for choosing a suitable regression method in a given application scenario. Regression methods were grouped in four classes: variable selection, latent variables, penalized regression and ensemble methods. The framework was applied to three case studies: two based on simulated data and one with real data from a wine age prediction study. Improved results were obtained when the models prior assumptions, regarding sparsity and collinearity, matched the data generating mechanism.