Evaluation of volatile and chemical profile of sherry-like white wine Vernaccia di Oristano from Sardinia by comprehensive targeted and untargeted approach

This study investigated the chemical–physical and volatile profile of sherry-like white wines with different aging time from three different areas of Sardinia. Proximate chemical composition, CIELab color coordinates and absorbances at 280 and 420 nm were determined. Volatile compounds were determined by solid-phase microextraction (SPME) followed by gas chromatography coupled with a mass spectrometer (GC/MS) using a targeted and untargeted approach. Significant correlations among the age of the wine and most of the chemical physical parameters, such as alcohol content, total acidity, volatile acidity, glycerol and polyphenols were observed. A280 and A420 values were highest, as expected, in the oldest wines as well as an increase of a* low values for L* were found during aging. Fifty-nine volatile compounds were identified, among which ethyl acetate, amyl/iso-amyl alcohol, ethyl octanoate, benzaldehyde, ethyl decanoate and phenylethyl alcohol were predominant. The untargeted approach was able to discriminate wines according to their production area, and within each group, according to the year of production.


Introduction
Among the whole chemical components commonly detected in wine, volatile organic compounds (VOCs) represent a very small fraction that is nevertheless strongly involved in the wine quality and it is considered a key factor in wine consumer satisfaction. In addition to the VOCs deriving directly from grape, known as varietal compounds, a very important fraction of volatile molecules, that contribute to the sensory perception of wine flavour, are produced by yeast during the fermentative process [1]. The aging process is responsible for significant changes of the wine volatile composition due to enzymatic and physicochemical reactions such as oxidation and reduction, which occur during wine storage and aging. The complexity of the final wine bouquet depends on many factors [2,3]. Usually, the presence of oxygen during white wine aging must be carefully controlled but in Sherry and Sherry-like wines, their uniqueness is due to the aerobic process that occurs in these wines responsible for the oxidative metabolism of flor strains of Saccharomyces cerevisiae [4][5][6].
Vernaccia di Oristano is an oxidized sherry-like white wine produced in Sardinia. Since 1971 this wine is disciplined by specific guidelines and protected by the Controlled Origin Designation (DOC), leading to four types of wine: Vernaccia di Oristano", "Vernaccia di Oristano" Superior, "Vernaccia di Oristano" Reserve and "Vernaccia di "Oristano" fortified.
As for the classical production of sherry and sherry-like wines, Vernaccia di Oristano is subjected to different biological and chemical processes that impact on its chemical, aromatic and sensory characteristics [7]. In fact, this wine undergoes biological aging, characterized from the presence of a natural flor velum which protects wine from any contact with air. The yeasts responsible for this velum formation are Saccharomyces cerevisiae flor yeasts, which, at the end of the alcoholic fermentation, when the wine is depleted from sugar and nitrogen compounds shift their fermentative metabolism to oxidative and create a multicellular aggregate on the wine surface [5]. During biological aging, usually performed inside chestnut barrels, the flor velum generates several compounds among which the volatile components play a key role in the sensory characteristics of the final wine [4].
Sherry-like white wine Vernaccia di Oristano has been the target of few papers and in literature the knowledge about its chemical characteristics is very scarce. Recently Tuberoso et al. [8] reported a chemical characterization of polar compounds as well as of undesirable biogenic amines of eleven Vernaccia di Oristano samples; the authors evidenced higher amounts of biogenic amines, if compared with classical white wines, but the content was similar to other oxidized and aged sherry-like samples. Furthermore, the authors also evidenced a good amount of phenolics, mainly represented by hydroxybenzoic and hydroxycinnamic acids, and flavonols. The VOCs compounds of Vernaccia di Oristano have been only partially explored. Galletti et al. [9] studied the volatile fraction before and after the inoculation of two different yeasts in Vernaccia di Oristano wine samples with the aim of better standardizing the ageing procedure. The authors evidenced the presence of ketones, aldehydes, alcohols, esters, acetals and acids which are typical classes of compounds already detected in sherry-like wines [10][11][12].
Moreover, the same authors [9,13] compared the volatile compounds produced during aging of Vernaccia di Oristano wine when inoculated with selected flor yeasts, with the volatile compounds produced during sterile aging. As expected, results showed that the flavor compounds concentration was dependent on the aging method; when flor yeasts were inoculated a bio-oxidation occurred while in sterile aging oxidative phenomena and evaporation losses were observed. At last, Corrias et al. [14] studied the effect of technological production process on pesticide residues in Vernaccia di Oristano samples, but no analysis of the VOC compounds was performed.
The aim of this work was then to characterize eighteen samples of wines Vernaccia di Oristano, from three different vocated areas in Sardinia (Cabras, Baratili S. Pietro and Zeddiani) by means of chemical composition, color, and volatile organic compounds (VOCs) to increase the knowledge about this sherry-like wine. Moreover, multivariate analytical techniques were applied to the results obtained with the aim to explore the data and to find any relationship among the parameters analyzed, the geographical area and the year of production. The proximate chemical composition was determined by WineScan™, while the color was determined both by tristimulus colorimetry using the CIELab space color and by spectrophotometry. The VOCs were determined both qualitatively and quantitatively by solid-phase microextraction (SPME) followed by gas chromatography coupled with a mass spectrometer (GC/MS). A targeted and untargeted approach was used with the aim to contribute to better understanding of this distinctive type of aging which leads to the production of sherry-like white wines.

Wine samples
Eighteen samples of wine 'Vernaccia di Oristano', aged from 1995 to 2019, were analyzed in this study ( Table 1). All samples came from the province of Oristano, situated in the central western Sardinia (Italy), to which three different vocated areas belong (Cabras, Baratili S. Pietro and Zeddiani). In these areas with peculiar climatic conditions, type of soil and presence of Saccharomyces cerevisiae flor yeasts, this wine can acquire sherry-like wines characteristics and labeled with the denomination of origin DOC. The Cabras area is characterized by alluvial and sandy soils, especially those close to the Tirso river, while Zeddiani and Baratili San Pietro are characterized by clay and sandy soils. All the three areas have a peculiar climate with hot and dry summers and rainy and moderate cold winters. In addition, the proximity to the sea gives these places a high degree of humidity.
Color measurements were carried out using a Minolta colorimeter (Minolta CR-300, Konica Minolta Sensing, Osaka, Japan), according to the recommendations of the International Commission on Illumination, with the illuminant D65 (daylight source) and 10° standard observer, and the results were expressed in accordance with the CIELab 1976 color space. The parameters calculated were a* (red/green values), b* (yellow/blue values), and L* (lightness). From the CIELab space were also calculated the parameters C* and h* (chroma and hue angle, respectively) [15]. All the analyses were performed on centrifuged wine samples for 15 min at 3000 rpm without any dilution. Spectrophotometric analysis (absorbance at 280 and 420 nm) was carried out by spectrophotometry (Hewlett Packard UV-Vis spectrophotometer Mod. 8453, Palo Alto, CA). The A 280 values were determined by diluting wine samples 1:10. All the measurements were carried out in triplicate.

SPME conditions for wine VOC analysis
VOC chemical composition was obtained by subjecting Vernaccia wines to solid-phase micro-extraction (SPME) coupled with gas chromatography according to a previously optimized method [16]. A 100 μm PDMS/DVB/CAR (Polydimethylsiloxane/Divinylbenzene/Carboxen) coated fiber (Supelco, Sigma Aldrich, St. Louis, MO, USA) was used to isolate the headspace volatile compounds. The fiber was preconditioned according to the manufacturer's instruction. 10 mL of wine and 100 µL of internal standard (3-octanol, 225 mg/L) were placed into a 20 mL SPME vial (75.5 × 22.5 mm) closed with a screw cap with Viton 1A septum. The fiber was conditioned for 5 min at 60 °C, then it was injected through the septum and suspended in the headspace. The fiber was exposed to the volatiles for 30 min; then it was retracted, removed from the vial, and placed immediately into the injector of the GC. Thermal desorption was carried out in the injector at a temperature of 250 °C for 5 min in splitless injection mode. The fiber underwent a further bake-out step for 5 min at 250 °C before and after each analysis.

Wine VOC gas chromatography-mass spectrometry (GC-MS) analysis
The volatile organic compounds were analyzed using a 7890 GC equipped with a Gerstel MPS autosampler coupled with an Agilent 7000C MSD detector according to Petretto et al. [16]. The chromatographic separation was performed using a VF-Wax column 60 m × 0.25 mm i.d., 0.5 µm film thickness (Agilent). The temperature program used was reported in Petretto et al. [16]. At the end of the column, the flow was split into an Agilent 7000C MSD detector. The data were analyzed using a Mass Hunter Workstation B.06.00 SP1. The identification of the individual components was performed by comparison against co-injected pure compounds and by matching the MS fragmentation patterns and retention indexes using the built-in libraries, literature data, or commercial mass spectral libraries (NIST/EPA/NIH 2008, HP1607 purchased from Agilent Technologies). A hydrocarbon mixture from C8-C23 was injected under the same HS-SPME/GC-MS conditions to obtain the linear retention indexes. The internal standard method (3-octanol) was performed to quantify the compounds in the headspace. The standards, where available, were accurately weighed and dissolved in 10 mL ethanol and then added to the internal standard to obtain five concentrations. Results were expressed as µg/L of wine.
In addition to the semi-quantitative and quantitative analyses, the GC-MS data were analyzed by untargeted approach using the XCMS online platform (https:// xcmso nline. scrip ps. edu) which allows the features detection, retention time alignment and preliminary statistics, as well as some useful statistical tools for the comparison of results. Briefly, raw chromatograms, once converted in.netCDF format were loaded into the online XCMS, by organizing them in three different database, namely Baratili, Zeddiani and Cabras, which represent the three geographical areas. A multi-group job was created with positive polarity and retention time in minutes as general parameters, "obiwardp" retention time correction with profStep (m/z) set to 1, and "centWave" method for features detection, with the following parameters: ppm, 100; signal/noise threshold, 6; peak width, 5 to 10; mzdiff, 0.01; prefilter peaks, 3; prefilter intensity, 100; noise, 0. The alignment parameters were as follows: bw, 10; minifrac, 0.5; mzwid, 0.25; minisamp, 1. Significant features were identified by Kruskal-Wallis statistical test (p < 0.01).

Statistical analyses
Statistical analysis was carried out using the statistical package XLSTAT 2020.1.2 for Windows 10 (Addinsoft).
Principal Component Analysis (PCA) was performed on chemical-physical data, followed by Pearson correlation (p < 0.05) to highlight any significant relationships between these data and the wine aging year. One-way ANOVA was performed on colorimetric data followed by Pearson correlation (p < 0.05). A Principal Component Analysis was then performed on the results from XCMS multigroup analysis, selecting the candidate features with p-value ≤ 0.01, using the Unscrambler X 10.5 software (Camo, Asa, Norway) using the different wine samples as scores and the significant features as loadings. Before to run the PCA analysis all the variables were weighed by 1/ (standard deviation) transformation.

Chemical physical composition of Vernaccia wines
The proximate chemical composition of samples is summarized in Fig. 1. As can be seen in the biplot there is a good correlation among the age of the wine and most of the chemical physical parameters. In fact, Pearson's correlation performed on chemical data confirmed significant correlations (p < 0.05) with the wine aging time for the following parameters: alcohol content (r = 0.78); total acidity (r = 0.79); lactic acid (r = 0.67); tartaric acid (r = − 0.51); total dry extract (r = 0.73); glycerol (r = 0.77); k (r = 0.74); total polyphenols (r = 0.71) ash (r = 0.72); SO 4 (r = 0.74) and CO 2 (r = − 0.66). More aged wines showed higher alcohol content. In fact, alcoholic grade ranged from 13.31% (v/v) for A sample to 23.53% (v/v) for C1 sample. The increase of alcoholic grade according to the age could be linked to the evaporation phenomena of water molecules through oak barrels. In fact, the balance between the rate of ethanol evaporation and of water is affected by the low environmental humidity always present in Vernaccia wineries during wine ageing, which usually does not exceed 40%. In fact, as reported in the literature, lower relative humidity increases the water evaporation, while higher humidity promotes the ethanol evaporation [17].
By contrast Martinez de la Ossa et al. [18] showed a decrease of ethanol concentration through aging, assuming a utilization of ethanol as a carbon source by yeasts. The same occurred for the total acidity content that increases during aging, probably due to the neoformation processes from flor yeasts. The same trend already found for ethanol concentration was also observed plotting the glycerol concentration versus ageing time which showed an increase of glycerol in older Vernaccia wines. Glycerol is a good carbon source for flor yeasts, and its content is higher when the flor yeast is not present [19]. Volatile acidity is considered a variable parameter that could increase due to the micro-oxygenation through the oak barrels that converts ethanol in acetaldehyde and then acetaldehyde in acetic acid, but it can also be consumed by flor yeasts for their  [19]. Another important parameter is the SO 4 content that increases during aging always due both to the evaporation phenomena of water molecules through oak barrels and to the sulfur dioxide oxidation. On the contrary the CO 2 content is higher in young wines, as expected. At last, with reference to the phenolic content, it was observed an increase during aging linked both to the evaporation process and to the release of phenolic substances from the barrels. During aging polyphenols undergo oxidation and condensation reactions that affect wine colour. In fact, during barrel aging, when the flor is not present, a slow oxidation of phenolic compounds occurs, causing browning phenomena [20]. In Vernaccia wine, the role of acetaldehyde in this phenomenon is crucial [21]. This molecule is always present in large quantities in Vernaccia because it is both synthesized from flor yeasts and produced in barrels because of oxidation phenomena occurring in wines when the barrels are drained. Acetaldehyde can form condensation products with tannins, particularly with monomeric and oligomeric forms, thus increasing wine brown colour and, at the same time, reducing wine astringency [20]. In addition, acetaldehyde is a precursor of other aroma compounds such as 1,1-diethoxyethane, sotolon and acetoin, contributing indirectly to the wine flavour [21]. With reference to the different geographical area, a separation among the three different areas can be observed, although it would be necessary to analyze more samples of the same year to put in evidence the role of the terroir. In fact, in this case, the separation observed seems to be more related to the year of production.
Color variations in Vernaccia wines were performed by spectrophotometric analysis at 280 nm and 420 nm of absorbance. The absorbance at 280 nm is often linked to the total phenol amount while the absorbance at 420 nm gives information about the oxidation state of phenols. A 280 ( Table 2) was higher in the oldest wines confirming our data related to the IP content (Fig. 1) and those reported in many papers about Sherry wines during aging [18,22]. High levels of absorbance at 420 nm were recorded for the oldest wine while low levels of absorbance were recorded for younger samples. Pearson's correlation coefficient for these data was − 0.83 (p > 0.05). Furthermore, browning can also be caused by a process that involves glyoxylic acid, formed by oxidation of tartaric acid in the wine in a reaction catalyzed by metals such as iron, and flavanols by means of a condensation-oxidation process [22]. At this point, glyoxylic acid can condense with flavanols similarly to acetaldehyde, giving rise to compounds that increase red hues in the wine. This reaction was confirmed by the results obtained from the CIELab parameters, where it was observed an increase of a* values during aging (coefficient correlation 0.67, p > 0.05) and more yellow hues in younger wines, whereas, with reference to lightness L* the oldest wines showed lower values than the young wines, as already reported in literature [11,15,20]. Pearson's correlation between the age of wine and the h* parameter confirmed, once again, what reported in  [15], where a decrease during aging was also observed in our wines due to the marked increase of red hues than the yellow hues (r = − 0.80).

Volatile profile of Vernaccia wines
Fifty-nine compounds were tentatively identified in Vernaccia samples (Table S1). Such compounds included aldehydes, esters, alcohols, ketones, and acids commonly detected in sherry like wines [9,[23][24][25]. Esters are commonly produced by the reaction between alcohol and organic acids. As reported in Table S1, the headspace of Vernaccia di Oristano evidenced the presence of esters obtained by reaction between methanol, ethanol, isoamyl alcohol and hexanol with acetic acid, fatty acids (C4-C16), benzoic acid and benzene acetic acid. In most samples, the presence of the two isomers of 5-Butyl-4-methyldihydro-2(3H)-furanone also known as "whiskey lactone" was detected. These compounds play a key role in the whiskey and sherry wines aroma and their presence in beverages is linked to the aging period in oak barrels [26] due to the extraction of their precursors by ethanol [25]. With reference to the acetals, that are tertiary aromas produced during wine aging, it was possible to identify two compounds, namely diethoxymethane and 1,1-diethoxy-2-methyl-propane, already well known as wine ageing markers [27]. Sotolon was not detected in our Vernaccia wines, probably due to its high boiling temperature which can reduce the sensitivity of the adopted headspace extraction technique [28]. In the headspace of Vernaccia wine samples acetoin, present only in traces, and two isomers of 2,3 butanediol, probably deriving from the acetoin reduction, were detected [4,21].
Although the great variability, as well as the great number of minor compounds, all the samples were characterized by compounds that appeared in all gas-chromatograms, namely ethyl acetate, amyl/iso-amyl alcohol (co-elution isomers), ethyl octanoate, benzaldehyde, ethyl decanoate and phenylethyl alcohol (Fig. 2). Similar results were also evidenced by Castro et al. [23] who compared two analytical extraction methods in the analysis of "fino" sherry wines. In addition to the above-mentioned volatiles, the authors also included in the main compounds of sherry wines several organic acids such as hexanoic and octanoic acids which were also detected in Vernaccia di Oristano as minor compounds. Anyway, the latter compounds are consumed as substrates by enzymes during vinification [18] therefore their concentration could be affected by the aging procedure as well as to the involved yeasts.
Quantitative analysis of the main compounds detected in the samples was carried out by the internal standard method and results are reported in Table 3. When appropriate, Odor Activity Values (OAVs) were calculated as the ratio between the concentration of the volatile compounds and their respective odor thresholds. Volatile compounds with an OAV > 1 are considered to play an essential role in the wine aroma [29]. Ethyl acetate is the main ester found in the headspace and its concentration ranged between 29 mg/L of sample T and 537.6 mg/L of sample 95. The encountered levels in these Vernaccia di Oristano samples resulted higher than those reported for the same cultivar in older works [9,13] but they are in line with other sherry like wines [11,25]. As reported in Duran-Guerrero et al. [4], a positive correlation between the ethanol concentration and ethyl acetate content was observed in Amontillado Sherry wines, up to concentration levels of hundreds of mg/L. Ethyl acetate concentration depends on the balance of activity of alcohol acetyl-transferases during the ester synthesis and on the activity of ester-hydrolases during the ester hydrolysis [30,31]. It is important to underline that yeasts responsible for the biological aging like flor yeasts tend to consume ethyl acetate due to the fatty acid metabolism, whereas this molecule increases during chemical aging because of the ethanol oxidation [32], and to the action of acetic acid bacteria during aging [33]. This compound is responsible for the fruity and flower notes of wine, but it is also responsible of wine acescency, especially when its concentration is above 80 mg/L [34]. Regarding the results obtained, ethyl acetate concentration exceeds the odor threshold (12.3 mg/L) reported in the literature [35] in all the wines analyzed, ranging from OAV > 2.4 for the sample T to OAV > 44.8 for the sample 95. However, the odor rejection threshold of ethyl acetate depends on the wine style and age, as reported in  Table 3 Quantitative analysis of volatiles in Vernaccia di Oristano Samples by HS-SPME-GC/MS analysis Results are reported by the mean of three replicates ± standard deviation and expressed as µg/L of wine *Co-elution isomers of 2-methyl-1-butanol and 3-methyl-1-butanol were quantified as equivalent of 2-methyl- 4.92 ± 0.0 the paper of Miranda et al. [33] where two panels found an odor rejection threshold for Madeira wines of 328 mg/L. In our case, only three wines, sample 95 1995, C2 1998 and C5 2004, exceeded this value, suggesting the possibility of bacterial contamination for them. Among esters ethyl butanoate and ethyl hexanoate, responsible for fruity aromas such as green apple, banana, strawberry, and floral notes, show a concentration higher than their respective odor threshold, 20 and 14 μg/L [15] with OAV > 1. With reference to the ethyl octanoate, except the sample C6, whereas the concentration is slightly higher than the odor threshold (5 μg/L), all the other samples have OAV ranging from 18.3 for sample C3 1998 to 126 for sample Co 2018, while ethyl decanoate values, that confer to the wine chemical and fatty odors were found to be lower than its odor threshold, 200 μg/L [15] with OAV < 1 for all the wines analyzed. In general, the concentration of ethyl esters detected in Vernaccia di Oristano samples is in line with other sherry like wines [23] as well as with other dry white wine samples [36].
Among alcohols, amyl/isoamyl alcohol (2-methyl-1butanol and 3-methyl-1-butanol in co-elution) resulted in the main components of this class of compounds (Table 3). Their concentration in our samples resulted slightly higher than those reported by Galletti et al. [9] in Vernaccia di Oristano samples but the OAV obtained are similar to the data obtained by Zea et al. [32] in Amontillado Sherry wines. Higher alcohols in wine are produced by yeasts through the amino acid metabolism (Erlich pathway) or directly from sugars and their concentration is affected both from grape cultivation condition, in particular its amino acids content, and from the yeast strains used. During biological aging the flor yeast, being in contact with the wine for long periods produces higher alcohol almost exclusively through the Erlich mechanism due to the absence of sugars during aging [37]. By the same biochemical pathway phenylalanine is the precursor of 2-phenyl ethanol, an important compound involved in the sensory characteristic of wine and responsible for rose and honey odours. The concentration of 2-phenyl ethanol in Vernaccia di Oristano samples ranged between 30.68 and 18.11 mg/L for Z3 and C5 Vernaccia samples, respectively. These values are in line with previous results on Vernaccia di Oristano wines [9,13] and other sherry type wines [11], whereas most of the samples exhibit a high odour activity (OAV > 5).
During aging, different volatile compounds, mainly esters and acetates, such as ethyl isovalerate (r = 0.92), isobutyl acetate (r = 0.84) and ethyl 2-methylbutanoate (r = 0.79) tend to increase their concentration. This increase could be linked to the esterification/hydrolysis balance favored by the high concentration of ethanol in older Vernaccia [15]. Despite of this, some esters such as ethyl hexanoate and ethyl octanoate (r = 0.61 and 0.69, respectively), tend to decrease, but this trend can be observed in the paper of Ferreira et al. [38]. Octanoic acid (r = 0.66) and decanoic acid (r = 0.60) tend to decrease. Furfural increases during aging (r = 0.76) so the diethyl succinate (r = 0.83) and diethyl glutarate (r = 0.66) and the butyl ethyl succinate (r = 0.84) and ethyl 3-methylbutyl succinate (0.85) and the ethyl hydrogen succinate (r = 0.72). As reported in Jagatic Korenika et al. [39] and Voce et al. [40], diethyl succinate is considered as "ageing" ester because of its increase when wine stays in contact with yeasts cells, e.g. during the second fermentation in sparkling wines and/or when malolactic fermentation takes place.
The whole chemical composition of headspace is characterized by the combination of all volatiles which could come from different pathways, and it is mainly composed by varietal, fermentative and aging compounds. The untargeted metabolite profile commonly includes data on all compounds (identified and unknown), thus containing all the information related to the variability of the chemical composition. Currently, this approach has being used to analyze grapes and wines with the aim to find differences of terroir, climatic factors, cultivation systems, and winemaking process [41]. With reference to the untargeted approach on the volatile profile of Vernaccia wines, the multi-group XCMS online analysis allowed to obtain a set of 128 features, 73 of which were statistically significant (p < 0.01) (Fig. 3). The results of the Principal Component Analysis performed on the significant features, which play a key role in the discrimination of the samples, are reported in Fig. 4 (score plot). The first two principal components (PC1 + PC2) explained the 82% of the total variance, the 63% of which being explained by the PC1 alone. The PC1 − PC2 score plot (Fig. 4) showed that the information contained in the variables was able to discriminate the samples, mainly along the PC1, according to their production area, and highlighted a further discrimination within each group, according to the year of production. Thus, the untargeted approach performed on volatile data, supported by the automation given by a specifically intended software, confirmed to be a potent and user-friendly tool in studies involving the discrimination of wines based on both geographical origin and ageing. In addition, the Co sample, that is the only one aged in steel tank, resulted clearly separated from the other samples (Fig. 4), suggesting a possible role of the untargeted approach in distinguishing among winemaking processes.

Conclusions
A detailed chemical characterization of 18 samples of Sardinian sherry-like wines (Vernaccia di Oristano wine) has been reported. The qualitative analysis of the headspace shows the presence of aldehydes, esters, alcohols, ketones, acids that are typical compounds commonly detected in wines subjected to biological ageing. The quantitative analysis of the main VOCs, carried out by internal standard method, evidenced the predominance in the headspace of ethyl acetate, the mixture of amyl/isoamyl alcohol and phenyl ethyl alcohol, characteristic products of yeast metabolism, with high OAV. Both spectrophotometric and color CIELab analyses have confirmed the role of the aging time on the chemical and volatile profile of Vernaccia di Oristano sherry-like wines. Conversely by subjecting the raw gas-chromatogram data to multivariate analysis a discrimination according to the geographical area of cultivation of the vine and aging has been found. Further research will be needed to deeply investigate the potential of the untargeted approach to discriminate wine samples based on winemaking processes.
Funding Open access funding provided by Università degli Studi di Sassari within the CRUI-CARE Agreement.

Conflict of interest
The authors declare that they have no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.