Abstract
Background and goals Research into wines made from cold-hardy interspecific hybrids, which have been integral for the establishment of a grape and wine industry in the Upper Midwest, has largely lacked investigations into chemical composition and consumer perception. The goals of this project were to i) survey Iowa wine industry members on varieties they thought were best suited for premium wine production and ii) measure consumer hedonic scores and chemical composition of 20 commercial Midwest wines made from five varieties, selected based on the initial survey. Cluster analysis was performed on the sensory data and consumer preferences were correlated with wine composition.
Methods and key findings A survey of Iowa industry members identified five cold-hardy interspecific cultivars as growing best in the state: Brianna, Edelweiss, La Crescent, Marquette, and Frontenac. Chemical analyses of 20 commercial wines revealed that titratable acidity was generally greater than in Vitis vinifera wines. The most protein concentrations were observed in La Crescent and Frontenac wines. Consumers were clustered into five groups based on liking scores and the two largest segments showed a preference for wines with more residual sugar.
Conclusions and significance This is the first survey of chemical composition and consumer liking for Midwestern wines produced from cold-hardy interspecific hybrids. The high protein concentrations observed in red and white wines are notable, as these may affect tannin extraction and haze formation, respectively. Although average Iowa consumers prefer wines with substantial residual sugar (>20 g/L), there is evidence of multiple consumer segments with different residual sugar and varietal preferences.
Introduction
Due to their cold winters, Iowa and other Upper Midwest states (Minnesota, North Dakota, Nebraska, South Dakota, and Wisconsin) have no commercial plantings of Vitis vinifera cultivars. Commercial grape production in these and other cold climates relies on interspecific cultivars (Vitis spp.), including Vitis labruscana (also called “native”; e.g., Concord), French-American hybrids (e.g., Maréchal Foch), and newer cold-hardy cultivars (e.g., Marquette) (Pedneault et al. 2013, Atucha et al. 2018). Although reports exist on the composition of commercial wines produced from V. vinifera grapes and, to a lesser extent, from French-American hybrids, there are no surveys on the composition of commercial varietal wines produced from cold-hardy cultivars. Several authors have reported on basic wine chemistry and phenolic and volatile content of cold-hardy varietal wines (Slegers et al. 2015, Rice et al. 2017, Nicolle et al. 2019, Norton et al. 2020), but these studies examined wines produced in research settings, rather than commercial wines. Furthermore, few studies of consumer preference have examined interspecific hybrid wines and there is no published research of consumer liking of Midwestern wines. Reports in the literature concerning consumer preferences/attitudes to interspecific hybrid wines include one of consumer preference for New York State Seyval blanc (Berkey et al. 2011), one of consumer stigma toward Colorado-produced Chambourcin wine (Costanigro et al. 2021), and one from Brazil investigating consumer acceptance of wines made from several Brazilian-bred interspecific hybrid cultivars (Biasoto et al. 2014).
There are also no previous reports that correlated interspecific hybrid wine chemical composition with consumer hedonic scores, as has been reported for V. vinifera wines (Lund et al. 2009, Sáenz-Navajas et al. 2015, Wang et al. 2016). For example, a recent study on Australian wines produced from non-traditional varieties clustered consumers into three segments with distinct preferences for wine (Mezei et al. 2021). Overall preference and segment preferences were then correlated with wine chemistry. The existence of consumer segments for wines produced from cold-hardy interspecific hybrids and/or by commercial Midwestern wineries has not been explored.
Therefore, we surveyed Iowa grape and wine industry members to determine which cultivars grew best and were most representative of the state. Based on these results, a representative sample of 20 commercial wines made from five interspecific grape varieties were selected for chemical analysis and hedonic sensory evaluation. Consumers were then clustered to evaluate whether variety or composition correlated with hedonic scores for different consumer segments.
Materials and Methods
Industry survey
A survey was sent by email to Iowa grape and wine industry members in September 2019 to gather information about cultivar plantings, wine production, and general comments about the idea of an Iowa signature wine. The survey was open for two months, with several reminder emails sent to ask for participation. Questions and results are provided (Table 1 and Supplemental Table 1). This survey was approved for human subject participation by the Institutional Review Board at Iowa State University.
Selected survey results for Iowa grape and wine industry members. Fifty-one unique individuals responded.
Wine selection
Based on the industry survey results, five wine varieties (white: Brianna, Edelweiss, and La Crescent; red: Marquette and Frontenac) were selected as representative of the grapes grown and wines produced by the Iowa grape and wine industry. Fifteen commercially available wines of each variety were purchased and blindly tasted by five wine professionals with previous experience tasting these hybrid varieties. These wines were narrowed to four wines of each variety (for a total of 20 wines) that would be chemically analyzed and used for the consumer sensory evaluation portion of the study. A variety of styles (e.g., dry versus sweet, sparkling versus still) were chosen to expose consumers to a broad selection of Iowa wines. Wines were stored at room temperature in the dark for approximately two months until the consumer sensory evaluation was performed.
Chemical analysis
Chemical analysis was performed on all 20 wines used in the consumer sensory evaluation. All measurements were performed in duplicate. Residual sugar (RS) and acetic acid (AA) were measured using enzymatic assays (Megazyme). Percent alcohol (% alc.) was measured by near-infrared spectroscopy on an Alex-500 (Anton-Paar). pH was measured using an Orion 2-Star benchtop pH meter (Thermo-Fisher Scientific) and titratable acidity (TA) was measured using a Titrino plus automatic titrator (Metrohm). Glycerol was measured by a high-performance liquid chromatography-refractive index detector (HPLC-RID) method as described (Savits 2014). Tannin and total iron-reactive phenolics (IRP) were measured by the Adams-Harbertson assay as described (Heredia et al. 2006). Protein was measured using an ethanol precipitation, acid hydrolysis, amino acid quantification method modified from a recent report (Kassara et al. 2022). Following hydrolysis, the resulting amino acids were derivatized using the EZFaast kit (Phenomenex) according to the manufacturer’s instructions and validated to the manufacturer’s standards.
Sugar-related wine styles were assigned to the wines based on their RS in the following way: dry (<10 g/L), off-dry (10 to 19.9 g/L), semi-sweet (20 to 75 g/L), sweet (>75 g/L).
Consumer sensory evaluation
Untrained consumer participants were recruited through Iowa State University email lists by the Sensory Evaluation Center (Iowa State University, Department of Food Science & Human Nutrition). Inclusion criteria required that participants were at least 21-years-old, consumed wine three to five times/month, had no known allergies to sulfur dioxide or asthma, and were not knowingly pregnant, or planning to become pregnant during the study. All willing participants that met the inclusion criteria were selected. The consumer sensory evaluation occurred over five weeks in February/March 2021, with one varietal (four wines) presented each week. The number of participants and demographic information is presented (Supplemental Table 2). Each participant was assigned a four-digit code to anonymize subject data and allow researchers to link an individual’s responses during the multiple weeks of the study. There were 46 participants that completed all five sessions. The sensory evaluations were approved for human subject participation by the Iowa State University Institutional Review Board. Study participants gave written informed consent and were compensated for their participation ($5/week and an extra $25 for participating in all five weeks).
During the tasting session, participants were seated in individual sensory booths and presented with all four wines in plastic tumblers (30 mL each wine) coded with three-digit numbers along with napkin, consent form, pen, water, and spit cup with lid. All wines were presented in the booth at the beginning of the session due to COVID-19 pandemic protocols to avoid interaction between staff and participants during the session. Thirty mL of each wine was served at room temperature, except for sparkling wines, which were refrigerated overnight before each tasting and removed from the refrigerator immediately before the tasting session. Wines were presented in ascending RS, except for rosé wines, which were presented at the beginning of the red wine sessions (Marquette and Frontenac). This ordering was done to prevent sweetness carry over (Jackson 2008). Participants were instructed to evaluate each wine independently and were asked to rate their liking on a seven-point hedonic scale (extremely dislike to extremely like). Participants also ranked the four wines at the end of the session (data not shown). An additional preference question was asked during the final session to determine whether participants preferred the white or red wines in the study.
Statistics
Statistical analyses were conducted using JMP Pro version 15.0.0 (SAS Institute, Inc.) and GraphPad Prism version 9.2.0 (GraphPad Software). A one-way analysis of variance (ANOVA) was conducted to determine whether grape variety was a significant predictor of protein concentration (p < 0.05). A one-way ANOVA was also performed to determine if grape variety or wine style was a significant predictor of liking scores. Linear correlation statistics between hedonic liking and chemical analyses were calculated as Spearman’s r (p < 0.05) using GraphPad Prism.
Participants’ hedonic scores were used to rank the 20 wines for each participant, and hierarchical clustering (Ward’s method) on rankings was used to cluster participants. Rankings were not standardized and duplicate hedonic scores were assigned the same rank, with lower rankings taking into account the duplicate rankings above. For example, if a participant scored three wines as “Extremely Like” and two wines as “Like,” the “Extremely Like” wines were all given Rank 1, and the “Like” wines were both given Rank 4. The wine rankings for each of the 46 participants is available (Supplemental Table 3).
Results and Discussion
Industry survey and wine selection
A survey was sent to Iowa grape and wine industry members (Table 1). The survey polled industry members on the grape(s) they thought grew best in Iowa and the varietal wine(s) and style(s) that best represented Iowa. Respondents represented a range of vineyard/winery occupations and vineyard/winery sizes (Supplemental Data). There were 51 unique participants; however, the total number of responses for some questions was greater than 51 due to the “Check All That Apply” nature of some questions.
The five cultivars selected by industry members as growing best in Iowa were all newer, cold-hardy interspecific hybrids: Brianna (69%), Frontenac (65%), Marquette (49%), La Crescent (49%), and Edelweiss (39%). The top four of these cultivars were also selected by industry members as producing varietal wines that best represented Iowa, although Petite Pearl replaced Edelweiss for the fifth position, possibly because the former is a more recent release (2009) (Table 1). The cultivars Brianna, Frontenac, Marquette, La Crescent, and Edelweiss were also identified as having significant plantings from a previous industry survey (Tuck and Gartner 2014). These five cultivars were selected for further study based on their aforementioned importance, because all five are sufficiently cold-hardy to be grown throughout Iowa and because varietal wines of each are produced by several commercial wineries. Respondents gave mixed answers regarding the preferred sweetness for the best Iowa wine style, with semi-sweet white as the top answer (67%), followed by dry red (35%), semi-sweet red (35%), and dry white (29%) (Table 1). Therefore, we selected a large array of sweetness styles and included specialty styles like rosé, sparkling, and fortified.
Fifteen Iowa wines made from each of the five cultivars were purchased and tasted blindly by wine professionals with good knowledge of hybrid cultivars. Wines were scored individually (20-point scale), followed by discussion to identify a set of four wines made from each cultivar that were perceived as fault free and covered a range of styles (total = 20 wines). Compositional and hedonic data were determined for each wine (Table 2). Initial chemical testing revealed one wine had a chemical parameter greater than United States regulatory limits; it was exchanged for a similar wine (variety and style) from Nebraska.
Summary of consumer sensory data and chemical analyses for all wines used for the sensory evaluation. Wine data are presented in the order they were presented to consumer participants. Significantly different values (one-way analysis of variance, p < 0.05) for variety means are italicized, with the proceeding letters indicating the result of the Tukey’s post-hoc test. Total iron-reactive phenolics (IRP) and protein are presented as mean concentration ± standard deviation. RS, residual sugar; %abv, % alcohol by volume; TA, titratable acidity; nd, below limit of detection. The number of wines in each category of sweetness is listed in parentheses.
Chemical analyses
The chemical composition of the 20 commercial wines evaluated in this study was determined (Table 2). Dry V. vinifera wines typically have pH between 3.0 and 3.7 and TA between 5 and 8 g/L tartaric acid equivalents (Waterhouse et al. 2016). The pH values for cold-hardy interspecific hybrid varietal wines is within this range, but the upper end of the TA range is higher (5 to 15 g/L tartaric acid equivalents) (Watrelot et al. 2020). The wines used in this study all fell within these expected pH and high TA ranges.
As expected, TA and pH were inversely correlated (r = −0.543; Figure 1) across all wines. However, La Crescent wines had both higher pH and TA than the other white varietals. This may be due to La Crescent grapes having substantially more malic acid than Brianna or Edelweiss (0.075 molar equiv. versus 0.044 and 0.050 molar equiv., respectively; molar equiv. is assumed from the sum of malic acid and lactic acid molar equivalents as determined by HPLC-diode array detection; results not shown), and presumably also more minerals like potassium. Glycerol, a fermentation byproduct of yeast metabolism measured by HPLC-RID, ranged from 3.5 to 10.7 g/L among the varieties, compared to 7 to 10 g/L in V. vinifera (Waterhouse et al. 2016). In the hybrid wines of this study, however, it is unclear what impact glycerol may have on consumer sensory perception. A positive correlation was observed between glycerol and AA (r = 0.650; Figure 1). Higher concentrations of both compounds are typically observed in higher-gravity fermentations due to a yeast osmotic stress response (Waterhouse et al. 2016), although high concentrations of AA may also be produced by lactic or AA bacteria. Glycerol and % alcohol also correlated positively (r = 0.520; Figure 1), presumably because higher-alcohol wines were more likely to start with higher-gravity must with more osmotic stress.
Heat map for linear correlations, showing Spearman’s r (* denotes significance at p < 0.05) between chemical parameters and hedonic liking scores for ALL DATA.
Red wines in the study were analyzed for their phenolic content with the premise that these commercial wines would be low tannin, as reported previously (Nicolle et al. 2019, Norton et al. 2020, Watrelot 2021). Total IRPs in the red wines were generally lower than values reported for V. vinifera wines (median 723.7 mg/L catechin equiv. versus a V. vinifera range of 872 to 3005 mg/L catechin equiv.) (Heredia et al. 2006). Tannin concentrations measured by the Adams-Harbertson assay were below the limit of quantification and therefore much lower than reports of V. vinifera wines (Heredia et al. 2006) The one exception to this was Marquette 2, which had 1380 g/L IRP, comparable to some V. vinifera wines; however, the tannin concentration (150 g/L) was considerably lower than in V. vinifera wines (Heredia et al. 2006). As a caveat, the winemaking protocols for these commercial wines were unknown.
The protein concentration in each wine (Table 2) was determined by a recently developed method (Kassara et al. 2022) involving ethanol precipitation, dialysis, protein hydrolysis, and amino acid quantification by gas chromatography-mass spectrometry. Protein concentrations for Marquette and Frontenac wines have been reported previously using other methods (Nicolle et al. 2019, Norton et al. 2020). This is the first report of protein concentrations for the white cultivars Brianna, Edelweiss, and La Crescent. The greatest protein concentrations were observed in Frontenac and La Crescent wines (avg = 113 mg/L ± 46 and 101 mg/L ± 37, respectively). A one-way ANOVA revealed that variety is a significant effect for protein concentration using a post-hoc Tukey’s analysis (Table 2). The increased protein and reduced tannin of red wines produced from interspecific hybrids compared to V. vinifera wines has been reported previously (Springer and Sacks 2014, Springer et al. 2016). These observations may result from low initial tannin and high protein in the original hybrid grapes and may be further exaggerated by poor extraction of tannin during fermentation due to potential interactions between the two macromolecular classes, i.e., high protein content in hybrids results in lower tannin extractability (Springer et al. 2016).
Wines of Brianna, Edelweiss, and Marquette all had significantly less protein concentrations than the Frontenac and La Crescent wines (34 mg/L ± 6, 38 mg/L ± 20, and 46 mg/L ± 16, respectively). In white wines made from interspecific hybrid cultivars, the impact of high protein is not reported. Anecdotal reports suggest that wines made from hybrid varieties require more bentonite than anticipated for protein stabilization.
Consumer acceptability and cluster analysis
The consumer hedonic data were collected over five weeks (one session per week), with varying numbers of participants each week (60 to 75 participants; demographic data in Supplemental Data). Of the original panelists, 46 participants completed all five sessions. ALL DATA refers to the complete data set of all participants over the five weeks (1392 data points) and ALL 46 refers to the data set of the 46 participants that completed all five sessions (920 data points).
Using ALL DATA, we observed that both varietal and sugar level (dry, off-dry, semi-sweet, or sweet) were significant factors in the consumer scores (ANOVA, p < 0.0001 for both factors; Tables 3 and 4). Mean hedonic scores for the semi-sweet and sweet Brianna and Edelweiss wines were greater than others. Since Brianna and Edelweiss are described as having “grapey, foxy” aroma/flavor (Maniscalco 2012, https://mn-hardy.umn.edu/grapes/varieties), the findings suggest that Midwest consumers prefer “grapey, foxy” wines made from grapes with significant Vitis labrusca heritage. This grapey/foxy attribute is associated with several compounds, particularly methyl anthranilate and 2-amino-acetophenone (Acree et al. 1990). Neither the foxy sensory characteristic nor associated odorants were quantified in this current study, but previous research has demonstrated that consumers from California and Pennsylvania had different preferences for these specific aroma/flavor compounds (Perry et al. 2019), emphasizing the appropriateness of performing wine preference studies at regional levels.
Mean hedonic scores by variety for all participant scores (ALL DATA), the 46 participants who completed all weeks (ALL 46), and the clusters identified through Ward’s hierarchical cluster analysis. One-way analysis of variance (p < 0.05) was performed on ALL DATA to determine if variety is a factor. Mean comparisons for ALL DATA were performed using the Steel-Dwass method, with letters proceeding the values to indicate significantly different groups.
Mean hedonic scores by style for all participant scores (ALL DATA), the 46 participants who completed all weeks (ALL 46), and the clusters identified through Ward’s hierarchical cluster analysis. One-way analysis of variance (p < 0.05) was performed on ALL DATA to determine if variety is a factor. Mean comparisons for ALL DATA were performed using the Steel-Dwass method, with letters proceeding the values to indicate significantly different groups.
Correlations between overall consumer preferences (ALL DATA) and individual chemical components were examined (Figure 1). There was positive correlation between hedonic scores and RS (r = 0.292), and negative correlations between hedonic scores and alcohol, AA, glycerol, and IRP (r = −0.144, −0.181, −0.254, and −0.270, respectively). While these correlations were weak (r < |0.3|), they were all statistically significant (p < 0.05).
A subset of consumers was created using the 46 participants who completed all five sessions, thereby giving a complete data set for all wines. Using this data set (ALL 46) and the rank sums of hedonic scores for each wine, we determined that two wines were favored overall, and one was least favored. The sum of rankings (Supplemental Table 3) for Edelweiss 4 (sweet sparkling) and Frontenac 1 (sweet rosé) were considerably lower at 151 and 226, respectively, than all other ranking sums (overall range 151 to 680), indicating greater liking across many participants. Marquette 2 (dry) had the highest sum of rankings (680), indicating reduced liking across many participants.
To identify consumer preference segments, a hierarchical clustering analysis (Ward’s type) was performed on the ALL 46 data, resulting in five clusters. There was no correlation between cluster and participant age or gender (results not shown). Mean hedonic scores for each varietal and sweetness style were determined within each participant cluster (Tables 3 and 4). The ALL 46 data were compared to the ALL DATA using a χ2 test, which indicated that the ALL 46 data sufficiently represented the overall data set. Cluster hedonic means for each variety or sweetness style were determined (Tables 3 and 4). Cluster 2 possessed the same order of mean hedonic scores as ALL DATA (and ALL 46), however, with a greater overall range (3.38 to 5.49 versus 4.06 to 5.02 for variety, respectively).
To facilitate interpretation of participant cluster data, average concentrations of all chemical parameters were determined for the upper and lower quintiles of wines (based on hedonic scores) for each cluster (Table 5). For all clusters except Cluster 4, representing 89% of participants, the top quintile wines had high average RS (>40 g/L) and low alcohol (<12.5%). V. labrusca wines made up all four of Cluster 2’s top wines and none of the bottom wines. The other clusters showed less preference for labrusca-type wines. Cluster 4 participants (11%) preferred wines with lower sugar. Aside from RS (Cluster 2), glycerol (Cluster 2), and TA (Cluster 5), no other chemical parameter was significantly different between the top 20% and bottom 20% of wines within a cluster.
Description of the five clusters from hierarchical clustering analysis. Average chemical parameters from the top 20% and bottom 20% of wines are shown with standard deviations. Significantly different values from the top 20% to bottom 20% within the same cluster are shown in italics and underlined.
Results of the cluster analysis indicate the average consumer liked wines with substantial RS and made from labrusca-based varieties (Clusters 1, 2, and 3); however, some consumers preferred lower sugar or wines not made from labrusca varieties (Clusters 4 and 5, respectively). Our observation that some wine consumer segments prefer sweeter wines is both widely accepted in the non-technical wine literature (Thach 2021) and not well-substantiated in the technical literature. Studies of other fruits or fruit-derived products have emphasized the importance of sweetness perception in consumer liking (Shewfelt and Brückner 2000, Crisosto et al. 2005). Sweetness perception is based on both sugar content and acidity of fruit products. Several studies have shown that the greatest sugar concentrations are not always the most accepted or preferred by consumers when there is a correspondingly high acidity. Instead, there is an optimal sugar concentration that is in balance with other chemical parameters, particularly acidity, within the fruit or fruit-derived product. While there are some wine groups that promote the use of a sweetness scale (using a sugar-to-acid ratio), it is not widely applied throughout the wine industry. For the wines in this study, the sugar-to-acid ratio (Supplemental Table 4) had a low correlation to hedonic scores (r = 0.260, result not shown) and was less than the correlation of hedonic scores with RS (r = 0.292). Therefore, the sugar-to-acid ratio was assumed to not be a driver of consumer liking for these wines.
The current study considered only major chemical components associated with mouthfeel and taste and did not measure other important contributors to overall perception, including visual aspects of color hue and intensity, and flavor odorants. Several reports examined volatile compounds present in the five varieties we examined; however, these reports did not include a consumer sensory evaluation (Mansfield and Vickers 2009, Savits 2014, Rice et al. 2018, 2019). Future studies should include color analyses, volatile analyses, and descriptive sensory analyses, as has been reported for V. vinifera wines (Lund et al. 2009, Mezei et al. 2021).
Conclusions
Although the Midwestern U.S. wine industry has grown rapidly in recent years, and most sales occur through the tasting room to local consumers, little is known about these consumers and their preferences. The combination of Iowa consumer preference clustering and chemical analysis helps understanding of market segments for Iowa wines in Iowa and, potentially, the Midwest. Over half of the participants fell into Clusters 1 and 2, distinguished by preferences for semi-sweet and sweet styles of white varietals (Brianna and Edelweiss). This clearly supports the popular stereotype of Midwest consumers preferring sweet wines (Hammon 2021) and is corroborated by industry survey results that most top selling wines were sweet (54%, Supplemental Table 3). However, there is also evidence that other consumers (Cluster 4, 11%) preferred less sweet wines and gave low scores to the sweetest wines. This data shows grapegrowers and winemakers that a range of consumer preferences exist, which may help marketing initiatives and wine portfolio planning.
The chemical analysis performed for the study contributed to new knowledge for commercial versions of these five cold-hardy, interspecific wine cultivars. Notably, some wines had remarkably high protein concentrations and the red varieties had low tannin concentrations compared to V. vinifera wines. This suggests additional questions regarding protein stability in white wines and mouthfeel perceptions in red wines. Anecdotally, winemakers have commented on the increased rates of bentonite necessary to stabilize wines made from interspecific hybrids. Future research considerations should consider cultivar effect on protein concentrations and what other winemaking protocols could be employed to mitigate high protein (Springer et al. 2016, Nicolle et al. 2019, Norton et al. 2020). In terms of red wine, the high protein concentration is currently being considered as a factor in low tannin concentrations, which were again observed in this study. Further investigations into protein removal or disruption to increase tannin extraction/retention from these grapes is ongoing. The overall goal in increasing tannin concentration is to improve the mouthfeel of red wines through enhanced astringency perception.
This study aimed to increase knowledge about wines produced from interspecific hybrid cultivars through chemical testing and consumer sensory information. This information is useful for both researchers and grape and wine industry members to design future research experiments for grapegrowing and production protocols.
Supplemental Data
The following supplemental materials are available for this article in the Supplemental tab above:
Supplemental Table 1 Demographic and additional survey results from Iowa grape and wine industry members.
Supplemental Table 2 Demographic data by week for the consumer sensory evaluation (*demographics of participants that completed all five sessions). Demographics for participants that completed all five sessions (46 participants) is also presented.
Supplemental Table 3 Ranking of ALL 46 data for clustering analysis. Rank sums and averages are presented for each wine. The cluster assigned each panelist after the Ward’s cluster analysis was performed is also presented.
Supplemental Table 4 Sugar-to-acid ratios (residual sugar [RS] divided by titratable acidity [TA]) for all wines. B, Brianna; E, Edelweiss; L, La Crescent; M, Marquette; and F, Frontenac.
Footnotes
Funding for this project was provided by the Iowa Department of Agriculture and Land Stewardship through the Specialty Crop Block Grant Program. The authors thank the staff of the Midwest Grape and Wine Industry Institute for their help in wine selection and procurement, chemical analysis and the consumer liking studies; Dr. Aude Watrelot for help with wine selection and Watrelot’s lab for the analysis of tannin and iron-reactive phenolics content in wines; Leah Reever of the Sensory Evaluation Center for recruiting consumer participants; Audrey McCombs for statistical consultation, both the Protein Facility and the W.M. Keck Metabolomics Research Laboratory for protein analysis.
By downloading and/or receiving this article, you agree to the Disclaimer of Warranties and Liability. If you do not agree to the Disclaimers, do not download and/or accept this article.
- Received February 2022.
- Accepted November 2022.
- Published online March 2023
This is an open access article distributed under the CC BY 4.0 license.
References
Vol 74 Issue 1
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
