Identification of Potential Volatile Markers for Characterizing Argentine Wine Vinegars Based on their Production Process

Identification of potential volatile markers for characterizing Argentine wine vinegars based on their production process Marcelo Wagner a,b , Jorgelina Zaldarriaga Heredia a,b , M. Pilar Segura-Borrego c,* , M. Lourdes Morales c , Jos´ e M. Cami˜ na a,b , Silvana M. Azcarate a,b,** , Raquel M. Callej´ on c , Rocío Ríos-Reina c a Instituto de Ciencias de la Tierra y Ambientales de La Pampa (INCITAPCONICET-UNLPam), Santa Rosa, La Pampa CPA L6302EPA, Argentina b Facultad de Ciencias Exactas y Naturales (UNLPam), Santa Rosa, La Pampa CPA L6300CLB, Argentina c ´ Area de Nutrici´ on y Bromatología, Facultad de Farmacia, Universidad de Sevilla, Sevilla 41012, Spain ARTICLE INFO Keywords: Argentinian vinegars Volatile markers Variable selection HS-SPME-GC-MS ABSTRACT In Argentina, the production of quality wine vinegars has been barely exploited. Currently, most of the marketed vinegars are produced using rapid industrial fermentation systems obtaining vinegars with a high production rate, but with low quality and few organoleptic nuances. This work aims to study the volatile profile by headspace solid phase microextraction (HS-SPME) coupled to gas chromatography-mass spectrometry (GC–MS) and chemometrics to differentiate Argentinian wine vinegars according to their production process (industrial or traditional). A total of 92 volatile compounds were identified on the samples by using a strategy based on volatile profile processing using PARADISe® software. The complete volatile profile of all the samples was submitted to partial least squares discriminant analysis (PLS-DA) for the selection of variables with importance in the projection (VIPs). Thus, 37 volatile compounds with the potential to be markers of the manufacturing process were selected. The results obtained revealed, for the first time, the volatile profile of Argentine wine vinegars showing that the compound groups that, on average, exhibited higher relative areas were esters, acids, and alcohols. Thus, certain acids, aldehydes and terpenes were noticeably present in industrial wine vinegar. Conversely, the alcohols were strongly associated with traditional ones. In the case of esters, they were connected to the different types of wine vinegars. These compounds could be considered as potential volatile markers of quality and authenticity of these types of vinegars. Introduction Vinegar, a widely used condiment and food preservative worldwide, is obtained through a double fermentation process that involves both alcoholic and acetic fermentation of sugary and starchy substrates [1]. There are various methods and raw materials, such as malt, apple, rice, wine, among others, for vinegar production. Although for a long time, wine vinegar was considered a low-cost byproduct that arose spontaneously during wine production, in recent years, there has been a significant shift in its perception. Now, wine vinegar has become a prized culinary product, generating a growing demand for high-quality varieties [2,3]. The process of obtaining wine vinegar can be carried out through two different acetification methods: the surface culture system and the submerged culture system. Surface culture systems involve traditional methods using wooden barrels, where acetic acid bacteria come into direct contact with atmospheric oxygen at the air/acetifying liquid interface or adhere to supports such as wood shavings. This process is typically slow and can take several months. It was the first method employed in vinegar production, as in the Orleans, Luxembourg, or Schutzenbach methods [4]. Vinegars produced through this slow process have a complexity and flavor comparable to those of fine wines, consequently making them more expensive. In summary, this method yields select and diverse products with a strong sensory impact. However, its main disadvantages include the extensive production time, high investment and installation costs, elevated final prices, and a lack of * Corresponding author. ** Corresponding author at: Instituto de Ciencias de la Tierra y Ambientales de La Pampa (INCITAPCONICET-UNLPam), Santa Rosa, La Pampa CPA L6302EPA, Argentina. E-mail addresses: [email protected] (M.P. Segura-Borrego), [email protected] (S.M. Azcarate). Contents lists available at ScienceDirect Talanta Open journal homepage: www.sciencedirect.com/journal/talanta-open https://doi.org/10.1016/j.talo.2024.100370 Received 8 March 2024; Received in revised form 30 September 2024; Accepted 16 October 2024 Talanta Open 10 (2024) 100370 Available online 18 October 2024 2666-8319/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/bync-nd/4.0/ ). control in the process [5]. In contrast, the submerged culture method is based on a rapid and forced fermentation, where the acetification medium is continuously rinsed with atmospheric air or enriched with oxygen, under conditions that favor maximum transfer from the gaseous phase to a liquid phase. Acetic acid bacteria are fully immersed in the liquid broth undergoing acetification. This way, higher yields and acetification rates are achieved [6]. Most commercial vinegars intended for common household consumption in Argentina are produced using this method. In the submerged culture system, the main limiting factors include oxygen, acetic acid concentration, and the availability of ethanol. The primary disadvantage of this cultivation system is the loss of volatile aromatic compounds, which are removed along with the air exiting the fermenter [7]. The scent of each wine vinegar is one of the main indicators of quality and depends on the profile of volatile compounds [8]. The distinctive final volatile profile of a vinegar is defined by various elements, including the raw material used, the production process, and the aging period in wooden barrels [9]. Even though Argentina stands out as a significant wine producer, the efficient exploitation and profitable production of wine vinegars have been limited. In reality, to achieve high production rates with reduced costs, there is a primary reliance on rapid industrial fermentation in vinegar production, resulting in low-quality products with limited organoleptic complexities [10]. Furthermore, to our knowledge, the volatile profile of wine vinegars produced in Argentina has not been studied. Gas chromatography–smass spectrometry (GC–MS) has become the predominant technique for examining the volatile compounds present in vinegars. To carry out these analyses accurately, it is essential to perform a prior extraction to concentrate the volatile compounds. Various extraction techniques have been employed to examine the volatile composition of wine vinegars, such as dynamic headspace sampling (DHS), headspace Extraction (HSE) [11], Solid-phase Extraction (SPE) [12], headspace solid-phase microextraction (HS-SPME), stir bar sorptive extraction (SBSE), and headspace sorptive extraction (HSSE) [8]. Although all have demonstrated their usefulness in vinegar characterization, HS-SPME has been one of the most employed techniques for analyzing the volatile profile of wine vinegars. It offers significant benefits by eliminating extraction solvents and by simultaneously performing the extraction and concentration steps in a short period [13]. Additionally, given that wine vinegar constitutes a complex matrix and GC–MS analysis provides abundant information, addressing the treatment of the generated data poses a challenge. To overcome this, the combination of GC–MS with chemometric techniques such as PARAFAC2 has proven effective in mitigating issues associated with GC–MS analysis of complex mixtures, allowing for the maximum information of the volatile profile to distinguish between samples [14]. In this sense, various programs and algorithms have been developed that apply chemometric techniques to rapidly process complex data sets from GC–MS analysis, and one of them is the Deconvolution and Identification System known as PARADISe® [15]. This software allows PARAFAC2 to be applied to a complex GC–MS data set through a simple user interface, providing as a result the area and identification of the total volatile compounds present in the sample, as well as correcting errors typical of this type of data [13,16,17]. Thus, starting from the non-directed analysis that PARADISE® allows to perform, in which it has a complete characterization of each vinegar sample, those volatile compounds that serve as markers of the quality of the different vinegar types can be determined through the use of the Variable Importance for Projection (VIP). These VIPs, which are obtained by the application of PLS-DA algorithm, are useful for reducing data complexity and refining the final discriminant model, aiming to enhance discrimination success by eliminating features with limited influence on the model’s performance and retaining those considered significant [18]. In this context, this study aimed to perform an insight and compare, for the first time, the characteristic volatile profiles of different Argentine wine vinegars as a function of the production processes. For that, samples were analyzed by headspace-solid phase microextraction (HS-SPME) coupled with gas chromatography–mass spectrometry (GC–MS), followed by the use of advanced chemometric techniques, including PARAFAC2 for extracting volatile information, PLS-DA to discern patterns, and the use of the VIPs to finally select and identify potential volatile markers associated with each type of Argentine wine vinegar based on their production process. This comprehensive approach represents the first endeavor of its kind in characterizing and discriminating between Argentine wine vinegars, shedding light on their unique volatile profiles and production distinctions. Materials and methods Samples A total of 20 samples of Argentinean wine vinegars from various brands were acquired, sourced from local markets and wineries. Among them, 14 samples were industrial wine vinegars, while 6 were traditional wine vinegars. No specific information is provided on the raw material because most wine vinegars are made from a mixture of grape varieties. Likewise, those grapes can come from different regions so their origin cannot be specified either. For all samples, pH values ranged between 2.49 and 3.00 and total acidity was found within the range between 3.70 % and 6.82 % for TA. All samples were stored in the laboratory for over 24 h at a constant room temperature of 25 ±1 ◦C to stabilize their temperature. The measurement of all samples was conducted immediately after their opening. Chemicals The volatile compound standards for identification were procured from various commercial sources. Thus, for internal standard (IS), a solution of commercial standard 4-methyl-2-pentanol prepared in ethanol, sourced both from Merck (Darmstadt, Germany), and water from a Milli-Q purification system (Millipore, USA), was prepared according to Ríos-Reina et al., (2018) [19] and stored at −20 ◦C until application. Additionally, sodium chloride (NaCl) obtained from Sigma-Aldrich (Madrid, Spain) was employed to enhance salting out in the vial, facilitating the passage of volatile compounds into the vial’s headspace. Additionally, the linear retention index (LRI) was determined by injecting a series of straight-chain n-alkanes C10 to C40 (50 mg/L in n-hexane) purchased from Fluka (Madrid, Spain). Furthermore, the identification of volatile compounds utilized available standards obtained from various commercial sources including Merck (Darmstadt, Germany) and Sigma-Aldrich (Madrid, Spain). Headspace - solid phase microextraction and analysis by gas chromatography and mass spectrometry (HS-SPME-GC–MS) The fundamental conditions for sample extraction were based on a previously validated method for wines [20] and Spanish wine vinegars [19], which were slightly adjusted to enhance the efficiency of the procedure. The process was as follows: 5 mL of sample, 1.65 g of NaCl, and 10 μ L of internal standard (IS) were introduced into a 20 mL glass vial, which was placed in the thermostat-controlled autosampler tray at 20 ◦C for HS-SPME sampling. Subsequently, the samples were incubated for 60 min at 62 ◦C with an agitation speed of 250 rpm. After incubation, volatile compounds were extracted from the headspace of the vial by exposing a 2 cm triple DVB/CAR/PDMS SPME fiber at 45 ◦C for 40 min and desorbed using the splitless mode with an injector temperature of 250 ◦C for 240 s. GC–MS analysis was conducted using an Agilent 8890 GC system coupled with an Agilent 5977B quadrupole mass spectrometer (Agilent, Santa Clara, CA, USA). A CP-Wax capillary column (60 m ×0.25 mm) with a film thickness of 0.25 μ m and helium as the carrier gas at a flow M. Wagner et al. Talanta Open 10 (2024) 100370 2 rate of 1 mL/min was employed. The oven temperature program was as follows: 35 ◦C for 5 min, then increased to 220 ◦C at 2.5 ◦C/min (held for 7 min). Duplicate analyses of the samples were performed, and blank runs using an empty glass vial were conducted after each sample analysis. All data were recorded using MS ChemStation software (Agilent). Data processing, identification of volatile compounds and statistical analysis The analysis of chromatographic profiles proceeded in several steps. Initially, Parallel Factor Analysis (PARAFAC2) was employed through the PARADISe® software to integrate the areas of volatile compounds and identify them through the use of the NIST database. Subsequently, a classification analysis was carried out using partial least squares discriminant analysis (PLS-DA), utilizing the projection importance variables (VIPs) obtained for variable selection. PARADISe® emerges as a powerful methodology for analyzing complex chromatographic data based on PARAFAC2. This methodology effectively performs simultaneous deconvolution of pure mass spectra of peaks and co-eluting compounds, robustly calculating peak areas for all samples. It also facilitates peak identification using resolved mass spectra and the NIST MS library [15]. Additionally, it demonstrates efficacy in baseline and noise correction, providing the opportunity to identify minor compounds in complex volatile profiles, such as those in fortified wines, while streamlining statistical analyses. The use of this software involved several steps. Initially, chromatographic data were converted to net. CDF format and exported to AIA format using MSD ChemStation (version F.01.01.2317). Then, a selection of intervals is needed. In this study, a total of 176 intervals were selected along the chromatogram to run the modeling of each interval using PARAFAC2, setting the maximum number of components at 8 per interval and applying the non-negativity constraint. Careful optimization of two parameters, the percentage of fit and the percentage of core-consistency, was crucial to selecting the correct number of components and verifying the model’s validity. After modeling, a report with the areas of the 104 resolved peaks and a tentative identification of each one, made by comparing the resolved mass spectra with the database of the NIST MS library, was generated. Then, identification was assessed after a series of alkanes standards injections and the calculation of the LRIs of each compound, which were then compared with the NIST spectrum library (version 2.0), the LRI of reference standards, when available, and LRIs reported in the literature. Then, after identification, those peaks that were identified as contaminants and noise were eliminated, resting 92 volatile compounds. Finally, peak area values were normalized by dividing them by the internal standard area value, i.e., obtaining relative areas (RA). In this study, peak quantification was not considered, as the goal was an undirected approach for sample comparison and differentiation, adopting a somewhat directed approach in selecting relevant peaks. The next step involved reducing the number of volatile to select those more important for each kind of wine vinegar. Thus, since the purpose was to differentiate wine vinegars based on their production process and select potential volatile markers, a data reduction and compound selection strategy based on the selection of projection importance variables (VIPs) using partial least squares discriminant analysis (PLS-DA) was implemented. Thus, PLS-DA was not developed for a classification approach due to the limited number of samples, but it was applied to study variables with importance in the prediction (VIP) for data reduction and selection. Sequential PLS-DA classification models were developed, selecting volatile compounds with a high VIP value in each, considering those with a VIP ≥1 as relevant [21], following the strategy used by different authors [22–24]. The models were built and optimized for all samples by selecting the best pre-treatment and the optimal number of latent variables (LV) to retain, leading to the lowest classification error in cross-validated Venetian blind cross-validation. Additionally, significant differences between data were determined using analysis of variance (ANOVA) followed by a post hoc comparison test (Tukey’s test) using INFOSTAT software (FCA, National University of C´ ordoba, Argentina). PLS-DA was performed using PLS Toolbox 7.9.5 (Eigenvector Research Inc., Wenatchee, WA, USA) working in a MATLAB 2016a environment (Mathworks). Before any modeling, data were autoscaled. Heatmap analysis was constructed using MetaboAnalyst 4.0 software (Xia Lab, McGill University). Results and discussion Characterization and comparison of the volatile profile of the different types of Argentinean wine vinegars In this study, the employed methodology allowed for the determination of 92 volatile compounds in Argentine wine vinegars. Tables S1 and S2 (Supplementary material) present the means and standard deviations of the relative areas (RA) for each volatile compound in each wine vinegar. Regarding the identification of these compounds, 42 of them were confirmed by standard mass spectrum and LRI; and 18 were tentatively identified (TI). In addition, there were 32 compounds with an unconfirmed identification (only mass spectrum matched with the library). The volatile compounds with the highest RA for all samples were acetic acid and ethyl acetate. On the one hand, the concentration of acetic acid serves as a frequent reference for evaluating the total acid content, becoming a crucial metric for appreciating the qualities of the final products [25]. On the other hand, previous studies on the aromatic profile of vinegars indicated that ethyl acetate was one of the volatile compounds that exerted a strong influence on the final sensory profile of wine vinegars. This compound is known for its fruity notes such as apple, pear, and grape, as well as its floral undertones [2]. The content of ethyl acetate depends almost exclusively on the alcohol content and acidity of vinegar [26]. In industrial vinegars, which have a small amount of residual ethanol, the presence of ethyl acetate is similar to that found in the base wines, while in quality or traditional wine vinegars, characterized by a higher residual ethanol amount, the amount of ethyl acetate may be higher. This compound has also been used to identify the degree of aging of a vinegar, because it suffers a concentration process during aging in wood barrels [27]. All these factors could explain the higher presence of ethyl acetate in some traditional vinegars (Table S1). Hence the stronger fruity fragrance and aroma of these vinegars compared to industrially manufactured ones [28]. There are many tasters who love the aroma of ethyl acetate and rate vinegars that smell strongly of this compound as higher quality [29,30]. Other compounds that exhibited a significant RA in the Argentinian wine vinegars included isovaleric acid, octanoic acid, ethanol, 2-phenyl- ethanol, 2-methylbutyl and isoamyl acetate. Within them, isovaleric acid, which represents one of the most abundant acids in other studied vinegars [8,9,19,31] and provides distinctive spicy notes to vinegars [32], showed a higher presence in the samples of vinegars produced industrially (Table 1). Regarding octanoic acid, a contributor to fatty, dry, and dairy aromas [33] that lends a unique “floral-sweaty” aroma to vinegar samples, and which was found as one of the compounds found in higher proportion in many wine vinegars [9,31], showed, on average, higher RA in industrually produced samples. Regarding alcohols, ethanol and 2-phenylethanol showed in this study higher RA for traditional wine vinegars than for the industrial ones, being in agreement with other authors that showed a high presence of these two compounds in other traditional or high-quality wine vinegars [8,19,31,34]. Phenylethyl alcohol, according to the literature, has been described that the scent of phenylethyl alcohol is pink and floral, which could contribute to a pleasant sensory appreciation [35]. Concerning acetates, isoamyl acetate showed higher RA in the traditional samples analyzed in this study than in the industrially produced samples. This compound was one of the compounds with the highest RA in the M. Wagner et al. Talanta Open 10 (2024) 100370 3 samples, been in agreement with Ríos-Reina et al.[19]. It has been shown to lend a strong fruity aroma to vinegar, and it was one of the odor-active compounds in the five wine vinegars detected and described by the sniffing panel [31]. Furthermore, it has been reported that this compound is also a potentially significant contributor to the flavor of aged vinegar [34]. 2-Methylbutyl acetate, one of the acetic esters found in higher amount in other vinegars [31], and generally associated with banana and fruity descriptors [36], was found in similar quantities in both traditional and industrial vinegars. Moreover, the identified volatile compounds were also categorized into ten groups based on their chemical families: acids, alcohols, aldehydes, esters, hydrocarbons, ketones, lactones, terpenes, volatile phenols, and others. The average of the total values of RA for each chemical group is shown in Fig. 1. As observed, esters were the chemical family that exhibited the highest RA for both traditional and industrial vinegars. Esters are essential ingredients in vinegars, and most esters Table 1 Mean, standard deviations of the relative areas (RA) and score values of the variable importance in projection (VIP) index for each elaboration process by the “37- selected-compounds” PLS-DA classification model. Industrial Traditional Identification number Volatile compounds LRI ID VIP Scores for Y1 Mean RA ±SD Mean RA ±SD 1 Ethanol 934 A 1.41 3.65 3.67 18.08 33.39 2 2.2-Dimethyl-1.3-butanediol 1058 C 1.65 0.03 0.03 0.11 0.16 3 Isoamyl acetate 1111 A 1.01 5.82 5.98 13.27 18.67 4 3-Methyl-1-butanol 1211 A 1.17 1.30 1.15 3.76 6.28 5 Isoamyl butyrate 1293 B 1 7.51 0.19 0.33 0.68 0.99 6 Ethyl lactate 1340 A 1.26 0.06 0.09 0.15 0.20 7 1-Hexanol 1357 A 1.46 0.02 0.02 0.17 0.36 8 2.3-Butanedyl diacetate I 1375 C 1.23 0.16 0.10 0.09 0.06 9 2-Nonanone 1387 A 1.45 0.01 0.01 0.03 0.04 10 3-Ethoxy-1-propanol 1384 B 1 1.74 0.01 0.01 0.01 0.01 11 Tetradecane 1395 B 1 1.24 0.03 0.03 0.06 0.07 12 cis-Linalool oxide 1450 B 1.41 0.39 0.51 0.09 0.04 13 3-Furaldehyde 1459 C 1.36 0.31 0.38 0.08 0.05 14 2.3-Butanedyl diacetate II 1479 B 1 1.00 0.03 0.02 0.04 0.06 15 trans-Linalool oxide 1478 B 1 1.53 0.16 0.20 0.04 0.01 16 2-Ethylhexan-1-ol 1492 A 1.33 0.14 0.07 0.23 0.23 17 1.2-Propylene diacetate 1506 B 1 1.43 0.02 0.02 0.01 0.01 18 Ethyl-2-hydroxycaproate 1516 B 1 1.23 0.04 0.03 0.08 0.11 19 Benzaldehyde 1527 A 1.46 0.51 0.51 0.25 0.15 20 Propanoic acid 1544 A 1.02 0.13 0.11 0.08 0.08 21 [S-(R*.R*)]-2.3-Butanediol II 1590 C 1.72 0.08 0.05 0.19 0.26 22 Hexadecane 1596 A 1.33 0.05 0.09 0.12 0.22 23 Hotrienol 1608 B 1 1.87 0.05 0.09 0.01 0.01 24 1.3-Propylene Diacetate I 1659 B 1 1.13 0.27 0.24 0.32 0.38 25 Isovaleric acid 1676 A 2.30 3.91 1.86 1.22 1.42 26 4-Propyl-benzaldehyde 1850 C 2.99 0.43 0.23 0.24 0.12 27 3-Hydroxy-2.4.4-trimethylpentyl 2-methylpropanoate 1885 C 3.63 0.94 0.68 0.36 0.33 28 Hexanoic acid 1863 A 1.54 0.06 0.03 0.08 0.05 29 α -Phenylethyl butyrate 1982 B a 1.10 0.04 0.07 0.01 0.01 30 cis-Whiskey lactone 1994 A 1.26 0.01 0.03 0.09 0.20 31 Octanoic Acid 2056 A 1.32 3.51 2.83 1.66 1.10 32 4-Hexyl-2.5-dihydro-2.5-dioxo-3-furanacetic acid 2086 C 1.30 0.02 0.03 0.01 0.01 33 Ethyl cinnamate 2106 A 1.22 0.02 0.05 0.01 0.02 34 Ethyl 5-oxooxolane-2-carboxylate 2177 C 1.43 0.00 0.01 0.03 0.06 35 Methyl anthranilate 2178 C 1.24 0.03 0.04 0.02 0.02 36 n-Decanoic acid 2193 C 3.06 0.82 0.47 0.41 0.34 37 Dodecanoic acid 2347 A 2.16 0.07 0.02 0.05 0.03 LRI: Experimental Linear Retention Index values estimated by linear regression. ID: reliability of identification: A. mass spectrum and LRI agreed with standards; B. mass spectrum agreed with mass spectral data base and LRI agreed with the literature data:. a National Center for Biotechnology Information (2004); C. mass spectrum agreed with mass spectral data base: Mean RA : mean relative area values; SD: standard deviation. *37-selected-compounds to 2nd PLS-DA. Fig. 1. Bar graph of the average of the relative areas (RA) for each chemical group according to types of wine vinegar production. M. Wagner et al. Talanta Open 10 (2024) 100370 4 contribute a fruity or floral aroma [37]. In the analyzed samples, it is observed that the relative areas corresponding to traditional vinegars are 70 % higher than those of industrial vinegars. This may be because these compounds are synthesized during alcoholic fermentation, acetic fermentation, and aging stages, primarily from the condensation of alcohols and corresponding acids [38]. Within this group, the main volatile compounds contributing to the highest RA in both types of vinegars were ethyl acetate, isoamyl acetate, and phenethyl acetate. Of these mentioned compounds, ethyl acetate and isoamyl acetate exhibited higher average RA for vinegars of traditional origin. Conversely, phenylethyl acetate showed higher RA in industrially produced vinegars. Regarding the group of acids, the RA for traditional vinegars represented 25 % less than those in industrial vinegars (Fig. 1). The predominant acids in the analyzed samples were acetic acid and isovaleric acid, which showed a higher RA for industrial vinegars. This can be attributed to the formation, degradation, oxidation, reduction, and esterification/hydrolysis of acids during the manufacturing process [39]. Most acids in vinegar have fatty, pungent, rancid, and fruity flavors, which substantially impact the vinegarʼs aroma [40]. Regarding the chemical family of alcohols, Fig. 1 shows that the total alcohol amount for traditional wine vinegars was almost three times larger for industrial vinegars, which was not as expected due to it has been reported that alcohol concentration increased continuously during fermentation and then decreased abruptly during aging, being attributed to the formation of acetals through the process of evaporation or interaction with aldehydes to form acetals [39]. One reason could be that Argentinian traditional wine vinegars were not subjected to prolonged aging (not more than six months), so there was not a prominent decrease of alcohols. Another reason was that these traditional samples showed higher RA for ethanol and 2-phenylethanol, as was previously discussed, which could increase the summatory for alcohols. The aldehydes and ketones with carbonyl groups were mainly generated through the oxidation of alcohols, which were characterized by their creamy and fruity flavors in vinegar [34]. In the samples analyzed in this study, the RA corresponding to aldehydes in industrial vinegars were 67 % higher than those in traditional vinegars. On the other hand, ketones are formed during alcoholic and acetic acid fermentation as a by-product of the fermentation process. However, the accumulation of ketones has also been observed during aging [9]. This results in the RA of ketones in traditional vinegars being slightly higher (4 %) than in industrial vinegars. Lactones can originate from sugar degradation or the cyclization of hydroxy acids during fermentation, and they are believed to contribute to sweet aromas [41]. In the vinegar samples analyzed in this study, lactones represented a relative peak area value 65% higher in traditional vinegars than in industrial ones. Terpenes contribute to the aroma of vinegars by imparting floral notes [41]. The terpene content of a wine can be influenced by a wide range of winemaking factors, including grape variety and maturation process. These compounds are primarily present in grape skins, and their concentration varies from one grape variety to another [41]. This results in a significant variability of these compounds in the analyzed vinegar samples, as they are made from blends of grapes from different strains. The RA of industrial vinegars represents more than double those of traditional vinegars in Argentine vinegars. In a previously published work, which analyzed samples of orange vinegar, a clear decrease in most volatile compounds was observed during the acetification process of surface culture, mainly in the terpene group [42]. This is consistent with the results of this work. In another study, an opposite behavior was reported, that is, the presence of terpenes increased in vinegars made traditionally [43]. Volatile phenols are aromatic compounds associated with notes of smoke, leather, forest, and pepper [44]. In the analyzed wine vinegars, the RA of industrial vinegars are 65% larger than those of traditional ones. These results do not coincide with the samples of Spanish wine vinegars, in which a significantly higher amount of volatile terpenes was found in traditionally produced vinegars [43]. Another set of compounds analyzed was hydrocarbons. In the analyzed samples, the RA of traditional vinegars represented a 76 % increase compared to industrial vinegars. In food samples, hydrocarbons are generated through the decarboxylation of long-chain fatty acids and are linked to the geographical and botanical origin of the raw material. [45]. ANOVA was carried out to determine if it was possible to differentiate wine vinegars according to their production process, considering the RA of each compound. The only compound that allowed significant differences to be identified was isovaleric acid. This compound, which was present in higher amounts in industrially produced vinegars, is likely derived from the degradation of branched-chain amino acids present in the raw material [30]. The vinegars analyzed in this study exhibit considerable variability. This may be because vinegars in Argentina are made from blends of different varieties of wines, whether they are red, ros´ e, or white. This situation was also described by Ríos-Reina et al., (2020) [10]. In Argentina, there is no comprehensive regulation and strict control over vinegar production. The Codex Alimentarius Commission regulates the basic specifications that a wine vinegar must meet to be marketed in the country [46]. However, this law has not established a distinction between the different types of wine vinegar production, nor has it provided a statement that could serve as an indicator of quality. 3.2. Data reduction and selection of differential volatile compounds After exploring the full dataset with 92 volatile compounds, a feature or variable selection process was conducted using the VIP scores obtained by the development of a PLS-DA model, in order to select the possible volatile markers useful for vinegar differentiation according to the production process and therefore, their quality. The score and VIP plots of the PLS-DA models are presented in Fig. 2. This approach resulted in three different PLS-DA models. First, one PLS-DA classification model was developed based on 5 latent variables (LV), using data encompassing the 92 extracted compounds, hereafter referred to as PLSDA92 (Fig. 2A and D). Subsequently, aiming to identify the most effective and straightforward classification model with fewer variables and therefore select the variables most relevant for the differentiation, another PLS-DA model with 3 LVs was constructed based solely on those compounds identified as VIP compounds (VIP values >1), in the PLSDA92 model, hereafter referred to as PLS-DA37 (Fig. 2B and E). Finally, a third PLS-DA model was made to perform a new variable reduction based on the 16 volatile compounds with VIPs>1 selected in the PLS-DA37, hereafter referred to as PLS-DA16 (Fig. 2C and F). This sequential PLS-DA modeling strategy was implemented to assess how many variables or features could be removed from the model without diminishing the achieved discriminant performance by selecting the most relevant volatile compounds for the differentiation of the two types of vinegars. Thus, those compounds that did not contribute to the distinction between types of wine vinegars were not considered and only compounds relevant to differentiation were selected in the VIP selection process. Three PLS-DA models were sequentially performed for feature selection, the first one with the total number of variables, a second one with 37 volatile compounds selected as VIPs from the first model, and a third model performed with 16 volatile compounds selected as VIPs in the second model. By assessing the classification indices (Table S3, Supplementary Material), it could be observed that 16 volatile compounds were not enough for explaining the differences between both types of vinegars (Fig. 2C and F), while with only 37 volatile compounds, the classification indices improved. Therefore, 37 volatile compounds were finally selected as the possible markers for the differentiation of traditional and industrial Argentinean wine vinegars. As indicated, the study was not merely a characterization of each wine vinegar type individually but rather a differentiation of samples M. Wagner et al. Talanta Open 10 (2024) 100370 5 based on their production process. Table 1 displays the VIP values associated to each volatile compound. Among the 37 selected volatile compounds, isoamyl butyrate, isovaleric acid, 4-propyl-benzaldehyde, 3‑hydroxy-2,4,4-trimethylpentyl 2-methylpropanoate, n-decanoic acid and dodecanoic acid presented VIPs >2, which means that they are the volatile compounds with high importance in the differentiation among traditional and industrial wine Argentinian wine vinegars. Specifically, as could be observed in Tables S1 and S2, isoamyl butyrate and 3‑hydroxy-2,4,4-trimethylpentyl 2-methylpropanoate showed higher RA in the traditional wine vinegar samples. On the other hand, the rest of the compounds were, on average, in higher presence in industrially crafted wine vinegars. Isoamyl butyrate is a short-chain ester derived from alcohol and an acid. This compound is found in greater quantities in traditional vinegars, as these samples also had a higher amount of alcohols, which are its precursors [47]. 3-Hydroxy-2,4,4-trimethylpentyl-2-methylpropanoate is a complex ester with a characteristic mild odor and has only been reported in grapes by Ubeda et al., (2017) [20]. These authors referred to this ester as a varietal marker of grapes from the Chilean region [48]. In Argentina, the largest wine production is located in the area of the province of Mendoza. This area borders with the Chilean region mentioned above, sharing geographical characteristics of the land where the vine is grown. Although this ester is found in lower quantities than those reported in Chilean grapes, we could also be facing a varietal marker for Argentine grapes. Isovaleric acid, a short-chain acid, originates from the metabolism of 3-methyl-1-butanol by acetic acid bacteria. Acetic acid, besides being the most abundant acid found in all vinegars, is even the most predominant acid in industrially produced vinegars. This results in isovaleric acid exhibiting a similar behavior [32]. During fermentation, bacteria and yeast can synthesize longer-chain fatty acids from simpler precursors. For example, they can elongate shorter-chain fatty acids, such as acetic acid, to produce longer-chain fatty acids like decanoic acid and dodecanoic acid [49]. This is reflected in the higher presence of these acids in industrially produced vinegars, similar to acetic acid. In industrial vinegar production, more standardized and controlled processes are commonly employed compared to traditional production. This can favor the formation and retention of certain aromatic compounds, such as 4-propylbenzaldehyde [50]. The volatile profile of Argentine wine vinegars has not been described to date. However, there are some samples from other countries characterized in the literature. Wine vinegars of Spanish, Italian, and French origin have been studied by other authors, who have detailed the volatile compounds present in their samples. Of the 37 compounds selected in the current study as markers of samples with different production processes, some have not been described in the samples reported in the literature. Among other compounds, propanoic acid, butanoic acid, isovaleric acid, n-decanoic acid, dodecanoic acid, 3- methyl-1-butanol, and benzaldehyde have been identified in Spanish wine vinegars with different DOPs [8]. Liu et al. (2022) [37] worked with samples of different vinegars, including wine vinegars from Italy, Spain, and France. Ethanol, isoamyl acetate, and hexadecane were identified in the samples from all three origins. Hexanoic acid and ethyl cinnamate were only identified in the samples of Italian vinegars. Isovaleric acid was not detected in any sample. [S-(R*,R*)]-2,3-Butanediol was only detected in the Spanish origin samples. Benzaldehyde was detected in the Italian and French samples [37]. Comparison between each wine vinegar according to the possible volatile markers selected To emphasize the global influence of volatile composition in each vinegar variety and investigate how volatile composition can differentiate wine vinegars based on the production process, the values of the VIP index (Table 1) and the results obtained by the development of a heatmap (Fig. 3) encompassing the 37 relevant volatile compounds, were studied. A heatmap was created to highlight the global impact of volatile composition in each vinegar sample and explore how this composition can distinguish wine vinegars based on the production process (Fig. 3). For its construction, the 37 VIP volatile compounds selected in the PLSDA model were utilized. Heatmap analysis was conducted using Fig. 2. VIP scores for each compound including (A) all the compounds, (B) compounds after the first variable reduction, and (C) compounds after the second variable reduction. Score plots of the PLS-DA models with: (D) all the compounds, (E) compounds after the first variable reduction, and (F) compounds after the second variable reduction. The numbers (1 to 37) correspond to the volatile compounds described in Table 1. M. Wagner et al. Talanta Open 10 (2024) 100370 6 Pearson’s correlation coefficients. Traditional vinegar samples exhibit positive Pearson’s correlations, while industrial vinegars display negative correlations with most compounds. Ethyl-2-hydroxycaproate (LRI 1516), 4-propyl-benzaldehyde (LRI 1850), 3‑hydroxy-2,4,4-trimethylpentyl 2-methylpropanoate (LRI 1885), 2-ethylhexan-1-ol (LRI 1492), and 2,3-butanedyl diacetate I (LRI 1375) were the compounds showing the highest positive correlation with samples of traditional origin. This allowed a clear differentiation between both sample groups. Additionally, in the cluster analysis (using the Ward method), it was demonstrated how vinegars are distinctly separated based on their production type, indicating a similarity in volatile profiles for each of the discriminated groups. These compositional differences and similitudes can be clearly observed from PLS-DA 37 results (Fig. 4). Fig. 4A displays the sample distribution, through the scores, represented as a function of their production process for LV1 and LV3, which accounted for 20.8% and 15.8% of the variance, respectively. As can be seen in Fig. 4A, the score plot showed a grouping of the samples according to the production process that mainly occurs throughout LV1: traditional wine vinegars mainly associated with the negative side of LV1, and industrial wine vinegars related to the positive side of LV1. Furthermore, looking at the loading plot (Fig. 4B), it could be seen the variable distribution of the 37 volatile compounds selected as possible markers of each wine vinegar (summarized in Table 1), from which almost all chemical families were distinguished. From simultaneous evaluation of both plots, the volatile composition related to each group can be revealed. Thus, acids, such as propanoic acid, isovaleric acid, hexanoic acid, octanoic acid, 4-hexyl- 2,5-dihydro-2,5-dioxo-3-furanacetic acid, n-decanoic acid and dodecanoic acid, were noticeably present in industrial wine vinegar. Also, the aldehydes (3-furaldehyde, benzaldehyde and 4-propyl-benzaldehyde) and terpenes (cis-linalool oxide, trans-linalool oxide and hotrienol) were totally related to this kind of vinegars. Conversely, the alcohols, namely ethanol, 2,2-dimethyl-1.3-butanediol, 3-methyl-1-butanol, 1- hexanol, 3-ethoxy-1-propanol, 2-ethylhexan-1-ol, and [S-(R*.R*)]− Fig. 3. Hierarchical clustering and heatmap for each production process by Pearson correlation coefficient obtained by the 37 volatile compounds selected as VIPs. The numbers (1 to 37) correspond to the volatile compounds described in Table 1. M. Wagner et al. Talanta Open 10 (2024) 100370 7 2,3-butanediol II were strongly associated with traditional wine vinegars, together with others compounds as cis-whiskey lactone (a lactone) and hydrocarbures (tetradecane and hexadecane). In the case of esters, they were connected to the different types of wine vinegars. Traditional wine vinegars were mainly linked to isoamyl acetate ethyl lactate, ethyl- 2-hydroxycaproate, 3‑hydroxy-2,4,4-trimethylpentyl 2-methylpropa- noate and ethyl 5-oxooxolane-2-carboxylate; while industrial vinegars were higher relation with 2,3-butanedyl diacetate I, 1,2-propylene diacetate, 1,3-propylene diacetate I, α -phenylethyl butyrate, ethyl cinnamate and methyl anthranilate. Specifically, the results obtained by the study of the 37-VIP scores and of the heatmap confirm that the aforementioned compounds were the most relevant for the desired differentiation. All mentioned compounds, linked to each wine vinegar elaboration process according to the results of the statistical analyses performed, could be responsible for specific and distinct aromatic profiles for each elaboration form. As observed in the sensory analysis conducted in a previous study, these differences in volatile composition among vinegar types lead to variations in the overall aroma of these samples [31]. Conclusions In conclusion, this study has unveiled, for the first time, the volatile profile of Argentinian wine vinegars as well as the applicability of the differentiating volatile fingerprint to discriminate among various types of wine vinegars based on their production processes. By combining headspace solid-phase microextraction (HS-SPME) coupled with gas chromatography-mass spectrometry (GC–MS) and the application of chemometrics, notable results were achieved. Thus, through a strategy focused on a limited number of volatile compounds, the study successfully discriminated between industrial and traditional Argentinian wine vinegars. Out of the initially detected 92 volatile compounds, a subset of 37 compounds was identified through a multivariate approach as having the highest potential to differentiate among the production processes. The compound groups that, on average, exhibited higher RA were esters, acids, and alcohols. Isoamyl butyrate, isovaleric acid, 4-propyl-benzal- dehyde, 3‑hydroxy-2,4,4-trimethylpentyl 2-methylpropanoate, n-dec- anoic acid, and dodecanoic acid were key compounds in differentiation. Isoamyl butyrate and 3‑hydroxy-2,4,4-trimethylpentyl 2-methylpropa- noate showed higher RA in traditional wine vinegar samples. On the other hand, the rest of the mentioned compounds were, on average, more present in industrially crafted wine vinegars. The limitations of this study, such as the sample size, provide a wide range of possibilities for future research. Therefore, additional studies utilizing information on the odor thresholds of these compounds, along with their assessment through GC-olfactometry and sensory analysis, would be necessary to confirm this hypothesis. CRediT authorship contribution statement Marcelo Wagner: Writing – original draft, Investigation, Formal analysis. Jorgelina Zaldarriaga Heredia: Writing – original draft, Investigation, Formal analysis. M. Pilar Segura-Borrego: Writing – review & editing, Methodology, Investigation, Formal analysis, Data curation. M. Lourdes Morales: Writing – review & editing, Investigation, Data curation. Jos´ e M. Cami˜ na: Investigation, Supervision, Writing – review & editing. Silvana M. Azcarate: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Raquel M. Callej´ on: Writing – review & editing, Supervision, Conceptualization. Rocío Ríos-Reina: Writing – review & editing, Supervision, Methodology, Investigation, Funding acquisition, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments The authors express their gratitude to CONICET (Consejo Nacional de Investigaciones Científicas y T´ ecnicas, Project PIP2021-2023, grant 11220200100316CO) and ANPCyT (Agencia Nacional de Promoci´ on Científica y Tecnol´ ogica, Project PICT-2018-04496 and Project PICT- 2020-00462 of Argentina) for financial support. The authors are also grateful to the Spanish government for the infrastructure grant EQC2019-005785-P co-funded by the Ministry of Science, Innovation and University. They would also like to thank to the “VI Plan Propio de Investigaci´ on y Transferencia” of the University of Seville for Dra. Ríos- Reina current contract (Contract number USE-18644-Z). Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.talo.2024.100370. Data availability Data will be made available on request. Fig. 4. Scores (A) and loadings (B) plots of the PLS-DA model performed with the total of samples and the 37 VIP volatile compounds obtained from the second PLSDA model. M. Wagner et al. Talanta Open 10 (2024) 100370 8 References [1] M. Lastra-Mejías, E. Gonz´ alez-Flores, M. Izquierdo, J.C. Cancilla, J.S. Torrecilla, Cognitive chaos on spectrofluorometric data to quantitatively unmask adulterations of a PDO vinegar, Food Control 108 (2020) 106860, https://doi.org/ 10.1016/j.foodcont.2019.106860. [2] R. Ríos-Reina, S. Elcoroaristizabal, J.A. Oca˜ na-Gonz´ alez, D.L. García-Gonz´ alez, J. M. Amigo, R.M. Callej´ on, Characterization and authentication of Spanish PDO wine vinegars using multidimensional fluorescence and chemometrics, Food Chem. 230 (2017) 108–116, https://doi.org/10.1016/j.foodchem.2017.02.118. [3] P. Kadiroglu, FTIR spectroscopy for prediction of quality parameters and antimicrobial activity of commercial vinegars with chemometrics, J. Sci. Food Agric. 98 (2018) 4121–4127, https://doi.org/10.1002/jsfa.8929. [4] W. Tesfaye, M.L. Morales, M.C. Garcia-Parrilla, A.M. Troncoso, Improvement of wine vinegar elaboration and quality analysis: instrumental and human sensory evaluation, Food Rev. Int. 25 (2009) 142–156, https://doi.org/10.1080/ 87559120802682748. [5] W. Tesfaye, M.L. Morales, M.C. García-Parrilla, A.M. Troncoso, Wine vinegar: technology, authenticity and quality evaluation, Trends Food Sci. Technol. 13 (2002) 12–21, https://doi.org/10.1016/S0924-2244(02)00023-7. [6] J.M. Gomez, L. Romero, L. Caro, D. Cantero, Application of a gas recirculation system to industrial acetic fermentation processes, Biotechnol. Technol. 8 (1994) 711–716, https://doi.org/10.1007/BF00151474. [7] I. De Ory, L.E. Romero, D. Cantero, Optimum starting-up protocol of a pilot plant scale acetifier for vinegar production, J. Food Eng. 52 (2002) 31–37, https://doi. org/10.1016/S0260-8774(01)00082-6. [8] R. Ríos-Reina, M.P. Segura-Borrego, D.L. García-Gonz´ alez, M.L. Morales, R. M. Callej´ on, A comparative study of the volatile profile of wine vinegars with protected designation of origin by headspace stir bar sorptive extraction, Food Res. Int. 123 (2019) 298–310, https://doi.org/10.1016/j.foodres.2019.04.071. [9] F. Chinnici, E.D. Guerrero, F. Sonni, N. Natali, R.N. Marín, C. Riponi, Gas chromatography-mass spectrometry (GC-MS) characterization of volatile compounds in quality vinegars with protected European geographical indication, J. Agric. Food Chem. 57 (2009) 4784–4792, https://doi.org/10.1021/jf804005w. [10] R. Ríos-Reina, S.M. Azcarate, J.M. Cami˜ na, R.M. Callej´ on, Sensory and spectroscopic characterization of Argentinean wine and balsamic vinegars: a comparative study with European vinegars, Food Chem. 323 (2020) 126791, https://doi.org/10.1016/J.FOODCHEM.2020.126791. [11] C. Ubeda, R.M. Callej´ on, C. Hidalgo, M.J. Torija, A. Mas, A.M. Troncoso, M. L. Morales, Determination of major volatile compounds during the production of fruit vinegars by static headspace gas chromatography-mass spectrometry method, Food Res. Int. 44 (2011) 259–268, https://doi.org/10.1016/j. foodres.2010.10.025. [12] C. Ubeda, F. San-Juan, B. Concejero, R.M. Callej´ on, A.M. Troncoso, M.L. Morales, V. Ferreira, P. Hern´ andez, Glycosidically bound aroma compounds and impact odorants of four strawberry varieties, J. Agric. Food Chem. 60 (2012) 6095–6102. [13] C. Ubeda, D. Cortejosa, M.L. Morales, R.M. Callej´ on, R. Ríos-Reina, Determination of volatile compounds for the differentiation of PDO fortified wines with different ageing methods as a tool for controlling their authenticity, Food Res. Int. 173 (2023), https://doi.org/10.1016/j.foodres.2023.113320. [14] G. Baccolo, B. Quintanilla-Casas, S. Vichi, D. Augustijn, R. Bro, From untargeted chemical profiling to peak tables – A fully automated AI driven approach to untargeted GC-MS, TrAC - Trends Anal. Chem. 145 (2021) 116451, https://doi. org/10.1016/j.trac.2021.116451. [15] L.G. Johnsen, P.B. Skou, B. Khakimov, R. Bro, Gas chromatography – mass spectrometry data processing made easy, J. Chromatogr. A 1503 (2017) 57–64, https://doi.org/10.1016/j.chroma.2017.04.052. [16] C. Sales, T. Portol´ es, L.G. Johnsen, M. Danielsen, J. Beltran, Olive oil quality classification and measurement of its organoleptic attributes by untargeted GC–MS and multivariate statistical-based approach, Food Chem. 271 (2019) 488–496, https://doi.org/10.1016/j.foodchem.2018.07.200. [17] S. Pellacani, C. Durante, S. Celli, M. Mariani, A. Marchetti, M. Cocchi, L. Strani, Optimization of an analytical method based on SPME-Arrow and chemometrics for the characterization of the aroma profile of commercial bread, Chemom. Intell. Lab. Syst. 241 (2023) 104940, https://doi.org/10.1016/j.chemolab.2023.104940. [18] B. Mahieu, E.M. Qannari, B. Jaillais, Extension and significance testing of Variable Importance in Projection (VIP) indices in partial least squares regression and principal components analysis, Chemom. Intell. Lab. Syst. 242 (2023) 104986, https://doi.org/10.1016/j.chemolab.2023.104986. [19] R. Ríos-Reina, M.L. Morales, D.L. García-Gonz´ alez, J.M. Amigo, R.M. Callej´ on, Sampling methods for the study of volatile profile of PDO wine vinegars. A comparison using multivariate data analysis, Food Res. Int. 105 (2018) 880–896, https://doi.org/10.1016/j.foodres.2017.12.001. [20] C. Úbeda, R. del Barrio-Gal´ an, ´ A. Pe˜ na-Neira, M. Medel-Marabolí, E. Dur´ an- Guerrero, Location effects on the aromatic composition of monovarietal cv. Carignan wines, Am. J. Enol. Vitic. 68 (2017) 390–399, https://doi.org/10.5344/ ajev.2017.16086. [21] T. Mehmood, K.H. Liland, L. Snipen, S. Sæbø, A review of variable selection methods in partial least squares regression, Chemom. Intell. Lab. Syst. 118 (2012) 62–69, https://doi.org/10.1016/j.chemolab.2012.07.010. [22] R. Ríos-Reina, R. Aparicio-Ruiz, M.T. Morales, D.L. García-Gonz´ alez, Contribution of specific volatile markers to green and ripe fruity attributes in extra virgin olive oils studied with three analytical methods, Food Chem. 399 (2022), https://doi. org/10.1016/j.foodchem.2022.133942. [23] P. Rodríguez-Hern´ andez, M.J. Cardador, R. Ríos-Reina, J.M. S´ anchez-Carvajal, ´ A. Gal´ an-Rela˜ no, F. Jurado-Martos, I. Luque, L. Arce, J. G´ omez-Laguna, V. Rodríguez- Est´ evez, V. Gal´ an-Rela˜ no, Detection of Mycobacterium tuberculosis complex field infections in cattle using fecal volatile organic compound analysis through gas chromatography-ion mobility spectrometry combined with chemometrics, Microbiol. Spectr. 11 (2023) e0174323, doi:10.1128/spectrum.01743-23. [24] R. Rubio-S´ anchez, R. Ríos-Reina, C. Ubeda, Identification of volatile biomarkers of Trichomonas vaginalis infection in vaginal discharge and urine, Appl. Microbiol. Biotechnol. 107 (2023) 3057–3069, https://doi.org/10.1007/s00253-023-12484- 6. [25] M. Wagner, J.Z. Heredia, A. Montemerlo, D. Ortiz, J.M. Cami˜ na, M. Garrido, S. M. Azcarate, Multiparametric analysis and authentication of Argentinian vinegars from spectral sources, J. Food Compos. Anal. 125 (2023) 105801, https://doi.org/ 10.1016/j.jfca.2023.105801. [26] V. Palacios, M. Valc´ arcel, I. Caro, L. P´ erez, Chemical and biochemical transformations during the industrial process of sherry vinegar aging, J. Agric. Food Chem. 50 (2002) 4221–4225, https://doi.org/10.1021/jf020093z. [27] M. Cocchi, C. Durante, M. Grandi, D. Manzini, A. Marchetti, Three-way principal component analysis of the volatile fraction by HS-SPME/GC of aceto balsamico tradizionale of modena, Talanta 74 (2008) 547–554, https://doi.org/10.1016/j. talanta.2007.06.016. [28] C. Ubeda, R. Ríos-Reina, M. del P. Segura-Borrego, R. Callej´ on, M.L. Morales, Vinegar aroma profile and analysis, Adv. Vineg. Prod. (2019) 1–339, https://doi. org/10.1201/9781351208475-17. [29] R.M. Callej´ on, M.L. Morales, A.M. Troncoso, A.C. Ferreira Silva, Targeting key aromatic substances on the typical aroma of Sherry vinegar, J. Agric. Food Chem. 56 (2008) 6631–6639, https://doi.org/10.1021/jf703636e. [30] R.M. Callej´ on, M.L. Morales, A.C. Silva Ferreira, A.M. Troncoso, Defining the typical aroma of Sherry vinegar: sensory and chemical approach, J. Agric. Food Chem. 56 (2008) 8086–8095, https://doi.org/10.1021/jf800903n. [31] R. Ríos-Reina, M.P. Segura-Borrego, M.L. Morales, R.M. Callej´ on, Characterization of the aroma profile and key odorants of the Spanish PDO wine vinegars, Food Chem 311 (2020) 126012, https://doi.org/10.1016/j.foodchem.2019.126012. [32] M. Ugliano, E. Squillante, A. Genovese, L. Moio, Investigation on aroma compounds of Modena balsamic vinegars, in: Twenty-First Century (Proceedings of the 10th Weurman Flavour Research Symposium, 2003, pp. 733–736, https://doi. org/10.13140/RG.2.1.1306.5769. [33] S.M. Rocha, F. Rodrigues, P. Coutinho, I. Delgadillo, M.A. Coimbra, Volatile composition of Baga red wine: assessment of the identification of the would-be impact odourants, Anal. Chim. Acta 513 (2004) 257–262, https://doi.org/ 10.1016/j.aca.2003.10.009. [34] L. Zhang, Z. Qin, L. Zhang, Y. Jiang, J. Zhu, Dynamic changes of quality and flavor characterization of Zhejiang rosy vinegar during fermentation and aging based on untargeted metabolomics, Food Chem. 404 (2023) 134702, https://doi.org/ 10.1016/j.foodchem.2022.134702. [35] M. Etschmann, W. Bluemke, D. Sell, J. Schrader, Biotechnological production of 2- phenylethanol, Appl. Microbiol. Biotechnol. 59 (2002) 1–8, https://doi.org/ 10.1007/s00253-002-0992-x. [36] M. Cameleyre, G. Lytra, S. Tempere, J.C. Barbe, 2-Methylbutyl acetate in wines: enantiomeric distribution and sensory impact on red wine fruity aroma, Food Chem. 237 (2017) 364–371, https://doi.org/10.1016/j.foodchem.2017.05.093. [37] R.C. Liu, R. Li, Y. Wang, Z.T. Jiang, Analysis of volatile odor compounds and aroma properties of European vinegar by the ultra-fast gas chromatographic electronic nose, J. Food Compos. Anal. 112 (2022) 104673, https://doi.org/10.1016/j. jfca.2022.104673. [38] S. Li, P. Li, X. Liu, L. Luo, W. Lin, Bacterial dynamics and metabolite changes in solid-state acetic acid fermentation of Shanxi aged vinegar, Appl. Microbiol. Biotechnol. 100 (2016) 4395–4411, https://doi.org/10.1007/s00253-016-7284-3. [39] S. Al-Dalali, F. Zheng, B. Sun, T. Rahman, F. Chen, Tracking volatile flavor changes during two years of aging of Chinese vinegar by HS-SPME-GC-MS and GC-O, J. Food Compos. Anal. 106 (2022) 104295, https://doi.org/10.1016/j. jfca.2021.104295. [40] V.I. Petropulos, E. Bogeva, T. Stafilov, M. Stefova, B. Siegmund, N. Pabi, E. Lankmayr, Study of the influence of maceration time and oenological practices on the aroma profile of Vranec wines, Food Chem. 165 (2014) 506–514, https:// doi.org/10.1016/j.foodchem.2014.05.144. [41] A. Marrufo-Curtido, M.J. Cejudo-Bastante, E. Dur´ an-Guerrero, R. Castro-Mejías, R. Natera-Marín, F. Chinnici, C. García-Barroso, Characterization and differentiation of high quality vinegars by stir bar sorptive extraction coupled to gas chromatography-mass spectrometry (SBSE-GC-MS), LWT - Food Sci. Technol. 47 (2012) 332–341, https://doi.org/10.1016/j.lwt.2012.01.028. [42] C. Cejudo-Bastante, E. Dur´ an-Guerrero, C. García-Barroso, R. Castro-Mejías, Comparative study of submerged and surface culture acetification process for orange vinegar, J. Sci. Food Agric. 98 (2018) 1052–1060, https://doi.org/ 10.1002/jsfa.8554. [43] R.M. Callej´ on, W. Tesfaye, M.J. Torija, A. Mas, A.M. Troncoso, M.L. Morales, Volatile compounds in red wine vinegars obtained by submerged and surface acetification in different woods, Food Chem. 113 (2009) 1252–1259, https://doi. org/10.1016/j.foodchem.2008.08.027. [44] R. Castro Mejías, R. Natera Marín, M.D.V. García Moreno, C. García Barroso, Optimisation of headspace solid-phase microextraction for the analysis of volatile phenols in wine, J. Chromatogr. A 995 (2003) 11–20, https://doi.org/10.1016/ S0021-9673(03)00541-7. [45] A.M. Giuffr` e, N-alkanes and n-alkenes in virgin olive oil from calabria (South Italy): the effects of cultivar and harvest date, Foods 10 (2021), https://doi.org/ 10.3390/foods10020290. M. Wagner et al. Talanta Open 10 (2024) 100370 9