Sensory Determination of Peach and Nectarine Germplasms with Instrumental Analysis

The flavour and mouthfeel of peaches are crucial qualities of peach germplasm resources that significantly influence consumer preferences. In this study, we utilized 212 peach germplasm resources from the Nanjing Peach Resource Repository, National Fruit Germplasm facility, Jiangsu Academy of Agricultural Sciences as materials for sensory analysis, electronic nose analysis, and composition analysis via high-performance liquid chromatography (HPLC). In the sensory analysis, we divided 212 peach germplasms into three clusters based on hierarchical cluster analysis (d = 5). No.27, No.151, and No.46 emerged as the most representative of these clusters. The electronic nose was used to conduct an evaluation of the aroma profiles of the 212 peach germplasms, revealing that the primary distinguishing factors of peach aroma can be attributed to three sensors: W1S (methane), W1W (terpenes and organosulfur compounds), and W5S (hydrocarbons and aromatic compounds). The primary differences in the aromatic substances were characterized by sensors W2W (aromatic compounds, sulphur, and chlorine compounds) and W1C (aromatic benzene). The HPLC analysis indicated that the persistence of peach sensory characteristics was positively correlated with acids and sourness and negatively correlated with sweetness and the ratio of sugar to acids. The overall impression of the 212 peach germplasms revealed a negative correlation with acids, while a positive correlation was observed between the overall impression and the ratio of sugar to acids. Therefore, this study substantially contributes to the preliminary screening of the analysed specific characteristics of peach germplasms such as No.27, No.46, No.151, and No.211. These selections may provide valuable information for the potential creation of superior germplasm resources.


Introduction
Peaches (Prunus persica (L.) Batsch) are the most important fruits in the world.Peaches are native to China and widely planted over 10,000 hectares (15.02 million tons, FAOSTAT, 2020), accounting for over 60% of the global total of peaches and nectarines.The evaluation of peach germplasm resources is important for their protection and utilization and the full exploitation of their potential economic, social, and ecological value in China [1].The evaluation indexes used in this regard include morphological and biological characteristics, quality characteristics, stress resistances, and pest resistance.Recent studies have mainly focused on morphological and quality characteristics [2].
Peach fruit quality is composed of fruit appearance; internal, nutritional, and flavour quality; storage qualities; and stress tolerance [3].Sensory evaluation, electronic nose analysis, electronic tongue analysis, and GC-MS are used to assess fruit flavour and palate The 212 varieties of peaches/nectarines (Appendix A) analysed were harvested at the National Peach Germplasm Repository (Nanjing, China, 32 • 20 N, 118 • 52 E, located 11 m above sea level) from May to July 2022, with each variety represented by two trees.The test varieties were 5-year-old mature trees with a Y shape and a row spacing of 3 to 5 m, and conventional cultivation measures were employed.Samples (50 peaches/nectarines) of each variety were randomly collected upon reaching maturity (defined as the disappearance of green coloration on the bottom of the fruit) from the two trees [13,14].Samples were harvested before 11:00 a.m.A total of 30 fruits of each variety were immediately stored in a freezer (−80 • C) for E-nose and HPLC measurement.A total of 20 fruits of each variety were taken into the sensory room for evaluation the following afternoon.

Sensory Analysis
The quantitative descriptive analysis method was used to train the sensory panel and carry out the sensory analysis according to the Sensory Analysis-Methodology Paired comparison test (ISO, 5495:2005) [15].Different samples of peaches and nectarines were used to allow the panel members to recognize their qualitive characteristics.Based on their availability, health, and general food habits, ten candidates (postgraduate students and researchers from JAAS, comprising five females and five males) were pre-screened.They volunteered for the project.All subjects gave their informed consent for inclusion before they participated in the study.In this study, materials were taken from the National Peach Germplasm Repository, Nanjing, China.These materials are safe for sensory research.All the experimental procedures involving volunteers were conducted in accordance with the Declaration of Helsinki.Also, the volunteers had the ability to discriminate between products according to basic taste thresholds and to describe their perceptions.Panel descriptions and definitions for peach and nectarine attributes (flavour, palate, persistence, and overall impression) were developed through brainstorming and round-table consensus.
The next step was to develop the score sheet for the different attributes for analysis.Flavour was evaluated according to being fruity, floral, earthy, woody, caramel, nutty and herbaceous, or vegetative.Palate (aroma, sweetness, sourness, and bitterness and astringency), persistence, and over impression were evaluated by assigning categorical scores from 0 to 10 (0, no sensation; 0.1-1.9,absence of sensation; 2.0-4.9,weak; 5.0-7.9,moderate; 8.0-10, intense) [16].Each of them independently estimated 20 fruits of typical size and appearance and with similar ripeness per germplasm.Then, each panellist randomly chose two fruits from each variety for palate evaluation.The evaluated samples were not named until the end of the evaluation.The evaluations were judged at 10 a.m in individual booths under white light.

E-Nose Measurement for Fruit Aroma
In total, 15 fruits were randomly divided into 3 blocks for E-nose measurement and analysed in triplicate.Measurements were performed using a portable, commercially available E-nose (PEN3.5,Airsense Analytics GmbH, Schwerin, Germany).The sensor array of the PEN3.5 has ten metal oxide semiconductor-type chemical sensors (Table 1) capable of operating at high temperatures to permit the classification and identification of different volatile species.When the sensors were exposed to volatiles, the changes in the conductivity (G)-to-initial-conductivity (G0) ratio (G/G0, relative conductivity or response value) depended on G conductivity.The concentration of volatiles led to the deviation of G/G0 (greater than or less than 1).In this study, 10 g of flesh from each sample was placed into a 300 mL beaker covered with sealing film.The beaker was kept at 25 • C for 30 min for E-nose evaluation.The aroma measurement method used followed that reported by Yan et al. [17], with some modifications.We inserted the E-nose sampling needle into the sealing film and extracted the gas in the beaker for detection.The volatile gas was pumped over the sensors of the E-nose at a flow rate of 400 mL/min.The E-nose analyses were recorded over a range from 0 to 60 s.One to three stable signals (response values) were observed to be in the middle period of the data stabilization time [18].After each analysis, the sampling chamber was washed with an air-dried flow for 60 s.

Sensor Name
Sensor Sensitives

W1C
Sensitive to aromatic benzene W3C Sensitive to ammonia and aromatic compounds W5C Sensitive to nitrogen oxides W1S Sensitive to short-chain alkanes such as methane W2S Sensitive to alcohols, ethers, aldehydes, and ketones W3S Sensitive to long-chain alkanes W5S Sensitive to hydrocarbons and aromatic compounds W6S Sensitive to hydrogen W1W Sensitive to terpenes and organosulfur compounds W2W Sensitive to aromatic compounds and sulphur and chlorine compounds

HPLC Measurement for Sugars and Acids
A total of 15 fruits were randomly divided into 3 groups for HPLC measurement and analysed in triplicate.The method of HPLC measurement for sugars and acids followed was modified by Shen et al. [19].Using 5 mL of extracting solution (ethanol: 0.2% metaphoric acid v/v), 0.5 g of the flesh was ground and extracted using ultrasound waves for 1 h.The solution was centrifuged for 5 min at 10,000 rpm and 4 • C. The 0.8 mL supernatant was dried using a concentrator (Eppendorf, Concentrator plus, Sigma Aldrich, Saint Louis, MO, USA, 230 V/50-60 Hz) for 2 h.A total of 1.6 mL of ultrapure water was added to the concentrate to obtain sugar/organic acid solution.The samples were filtered through a 0.22 µm nylon syringe filter into an HPLC glass vial and capped tightly.Then, 0.5 µL portions of the samples were injected into an HPLC system with a 10 µm, 250 × 6 mm CARBOSep CHO-620 CA column (Transgenomic, Omaha, NE, USA) to detect soluble sugars at 80 • C using a differential refractive index detector.For the mobile phase, we used ultrapure water added at 0.5 mL min −1 .
A total of 0.5 µL of the samples was injected into an HPLC system with a 5 µm, 250 × 4.6 mm ZORBAX Eclipse XDB-C18 column (Agilent, Santa Clara, CA, USA) to analyse organic acids at 2 • C using a VWD UV detector (λ = 214 nm).The mobile phase was 0.02 mol/L of KH 2 PO 4 (pH 2.7) for 0.5 mL min −1 .Data were analysed using the Chemstation (Agilent) chromatography data system.KH2PO4 was used as a standard for each batch of samples.A total of 0.02 mol/L of KH 2 PO 4 (pH 2.7) (3.333, 1.667, 0.333, 0.167, and 0.033 g/L) was used to establish a standard curve for reporting organic acids.Soluble sugars and organic acids were identified by their retention times, and their concentrations were calculated in parallel to calibrate the internal ultrapure water and 0.02 mol/L of KH 2 PO 4 (pH 2.7) standard curves.

Statistical Analysis
Sensory and HPLC data were tested in terms of their average characteristics at maturity using IBM SPSS Statistics 22 (IBM, Armonk, NY, USA).Pearson correlation test (p < 0.05) was conducted to analyse the relationship between sensory indexes and sugars and acids.Principal component analysis (PCA) and hierarchical cluster analysis were performed using OriginPro 2021 (OriginLab Corporation, Northampton, MA, USA).Data related to E-nose were processed using the WinMuster program in PEN3.5 (Airsense Analytics GmbH, Schwerin, Germany) and OriginPro 2021.The main sensors used for main peach smell detection were analysed via loading analysis (LOA) using WinMuster software (Version 1.6.2.15-2011, Airsense Analytics, Schwerin, Germany).Principal component analysis (PCA) and hierarchical cluster analysis were performed using OriginPro 2021 based on main and aromatic substance sensors.
Persistence was described as the duration of flavours and aromas for a peach, ranging from none to long.As indicated in Figure 1B, No.164 showed no persistence, 39 germplasms had a short finish (0.1-1.9), and 112 peach fruits exhibited short to medium persistence (2.0-4.9),constituting the largest group in the study.No.38, No.47, and No.119 had long finishes (8.0-10.0).Overall impression encompassed final thoughts with a rating (score out of 10) and freestyle impressions, aiding in distinguishing the participants' favourite peaches.Germplasms were evaluated as excellent (8.0-10.0),good (5.0-7.9),medium (2.0-4.9), and poor (0.1-1.9), with 40 poor, 88 medium, 77 good, and 7 excellent germplasms.No.76 had a poor overall impression, while the highest scores for excellent impressions were those for No.209 and No.134.

Figure 1. (A)
. Palate scores of 212 peach germplasm resources: blue area corresponds to sweetness scores; orange area corresponds to sourness; grey area corresponds to astringency; and yellow area corresponds to bitterness.(B).Persistence and overall impression scores of 212 peach germplasm resources: blue area corresponds to persistence, and orange area corresponds to overall impression.
Persistence was described as the duration of flavours and aromas for a peach, ranging from none to long.As indicated in Figure 1B, No.164 showed no persistence, 39 germplasms had a short finish (0.1-1.9), and 112 peach fruits exhibited short to medium persistence (2.0-4.9),constituting the largest group in the study.No.38, No.47, and No.119 had long finishes (8.0-10.0).Overall impression encompassed final thoughts with a rating (score out of 10) and freestyle impressions, aiding in distinguishing the participants' favourite peaches.Germplasms were evaluated as excellent (8.0-10.0),good (5.0-7.9),medium (2.0-4.9), and poor (0.1-1.9), with 40 poor, 88 medium, 77 good, and 7 excellent germplasms.No.76 had a poor overall impression, while the highest scores for excellent impressions were those for No.209 and No.134.
'The Davis Wine Aroma Wheel' is an ideal tool for identifying and visualising the various tastes, smells, and aromatic qualities in most wines, regardless of grape variety.According to this wheel, the 212 peach germplasms contain special flavours such as fruity (passion fruit, lemon, citrus, apple, green apple, honeydew, mango, apricot, juicy honey peach, and yellow peach), floral (gardenia, jasmine, rose, and violet), herbaceous or vegetative (herb, grass, celery, carrot, and lotus seed), caramel (candy, popcorn, molasses, and honey), woody (smoky and rubber), nutty (peanut, cereal, and red bean), animal (butter/milk), and pungent (metal) as well as faults like being spoiled and exuding off-odours.'The Davis Wine Aroma Wheel' is an ideal tool for identifying and visualising the various tastes, smells, and aromatic qualities in most wines, regardless of grape variety.According to this wheel, the 212 peach germplasms contain special flavours such as fruity (passion fruit, lemon, citrus, apple, green apple, honeydew, mango, apricot, juicy honey peach, and yellow peach), floral (gardenia, jasmine, rose, and violet), herbaceous or vegetative (herb, grass, celery, carrot, and lotus seed), caramel (candy, popcorn, molasses, and honey), woody (smoky and rubber), nutty (peanut, cereal, and red bean), animal (butter/milk), and pungent (metal) as well as faults like being spoiled and exuding off-odours.Principal component analysis (PCA) revealed the influence of each sensory descriptor on the evaluation of peach germplasm resource (Figure 2A).The first principal component (PC1) accounted for 46.1% of the variance, with sweetness and overall impression contributing significantly.Sourness and persistence also had negative loadings for PC1.The variance contribution of the second principal component (PC2) was 19.2%, with astringency and bitterness serving as major contributors.Thus, PC1 reflected an enjoyable taste, while PC2 was associated with negative feedback.No.211 was characterized as sweet with a great impression, while No.101 was perceived as astringent and bitter with an unpleasant impression.No.119 had light sweetness, strong sourness, and a long finish, resulting in a harsh impression.Additionally, sweetness correlated closely with overall impression, as did sourness with persistence.Cluster analysis categorized the 212 germplasms into three groups (Figure 2B), with 127 in cluster 1, 13 in cluster 2, and 72 in cluster 3. No.27, No.151, and No.46 were the most representative of their respective clusters.

Instrumental Analysis
Response to Different Sensors Evaluated via E-Nose Based on the response values of 10 sensors, 212 peach germplasm samples were divided into four categories (d = 15) (Figure 3): 186 were assigned to Class I, 23 were assigned

Instrumental Analysis Response to Different Sensors Evaluated via E-Nose
Based on the response values of 10 sensors, 212 peach germplasm samples were divided into four categories (d = 15) (Figure 3): 186 were assigned to Class I, 23 were assigned to Class II, 2 were assigned to Class III, and 1 was assigned to Class IV.Consequently, No.82, No.91, and No.193 were deemed notable for their distinction from Class I and II.

Response to Different Sensors Evaluated via E-Nose
Based on the response values of 10 sensors, 212 peach germplasm samples were vided into four categories (d = 15) (Figure 3): 186 were assigned to Class I, 23 were assign to Class II, 2 were assigned to Class III, and 1 was assigned to Class IV.Consequent No.82, No.91, and No.193 were deemed notable for their distinction from Class I and II In Figure 4A, the W1W sensor had an influence on both the first and second princip components.The sensors W1S and W5S contributed to the first principal component.validate the significance of sensors W1W, W1S and W5S in differentiating the peach fr aromas, the three sensors were analysed in the current mode, as depicted in Figure 4 They collectively accounted for 99.99% of the total contribution, signifying a substant role in the discrimination of peach fruit aromas in the experiment.In Figure 4A, the W1W sensor had an influence on both the first and second principal components.The sensors W1S and W5S contributed to the first principal component.To validate the significance of sensors W1W, W1S and W5S in differentiating the peach fruit aromas, the three sensors were analysed in the current mode, as depicted in Figure 4B.They collectively accounted for 99.99% of the total contribution, signifying a substantial role in the discrimination of peach fruit aromas in the experiment.
Figure 5 exhibits the response values of No.193 to the various sensors of an electronic nose, with peak values around at 24s.A stable period for further analysis was determined to be from 35 s to 37 s.Among the ten sensors, W1S (short-chain alkanes) exhibited the greatest response, followed by W5S (hydrocarbons and aromatic compounds), W1W (terpenes and organosulfur compounds), W2S (alcohols, ethers, aldehydes, and ketones), W6S (hydrogen), W2W (aromatic components and sulphur and chlorine compounds), W5C (nitrogen oxides), W3C (ammonia and aromatic compounds), and W1C (aromatic benzene).The largest response value was observed for W1S (32.702 ± 0.015), followed by W5S (32.803 ± 0.067), W1W (24.373 ± 0.102), and W2S (8.789 ± 0.005).The other sensors had lower response values of around 1. The coefficients of variation for W5C, W2W, and W1W were 1.24%, 0.74%, 0.39%, and 0.42%, respectively.W1C and W3S displayed relatively minor coefficients of variation, amounting to 0.14% and 0.04%, respectively.Figure 5 exhibits the response values of No.193 to the various sensors of an electronic nose, with peak values around at 24s.A stable period for further analysis was determined to be from 35 s to 37 s.Among the ten sensors, W1S (short-chain alkanes) exhibited the greatest response, followed by W5S (hydrocarbons and aromatic compounds), W1W (terpenes and organosulfur compounds), W2S (alcohols, ethers, aldehydes, and ketones), W6S (hydrogen), W2W (aromatic components and sulphur and chlorine compounds), W5C (nitrogen oxides), W3C (ammonia and aromatic compounds), and W1C (aromatic benzene).The largest response value was observed for W1S (32.702 ± 0.015), followed by W5S (32.803 ± 0.067), W1W (24.373 ± 0.102), and W2S (8.789 ± 0.005).The other sensors had lower response values of around 1. The coefficients of variation for W5C, W2W, and W1W were 1.24%, 0.74%, 0.39%, and 0.42%, respectively.W1C and W3S displayed relatively minor coefficients of variation, amounting to 0.14% and 0.04%, respectively.response values of the three aroma sensors, germplasms in Class I responded notably to the sensor W1C, while germplasms in Class II responded significantly to sensor W2W.This result was consistent with the previous results for the coefficient of variation.It indicates that the differences in peach aromas predominantly arose from non-aromatic substances such as methane, nitrogen oxides, and hydrogen sulphide, with variations in aromatic substances primarily reflected in the response values of W2W and W1C.

Sugars and Acids Identified via HPLC
Figure 8A presents the sucrose, glucose, fructose, and sorbitol content of 212 germplasms determined via HPLC, with sucrose ranging from constituting 11.9% to 83.3% (from 4.61 mg/g to 118.28 mg/g) of the total sugar content, while sorbitol corresponded to the lowest content, ranging from 0.01 to 26.80 mg/g (from 0% to 18.6%), in the 212 germplasms.The glucose and fructose percentages were similar, spanning from 3.8% to 47.7%.No.158 had the lowest total sugar content at 25.86 mg/g, whereas the highest content was 172.61 mg/g, corresponding to No.73.In Figure 8B, the total acid content varies from 2.07 mg/g in No.150 to 33.61 mg/g in No.76.The quinic acid content ranges from 0.32 to 10.72 mg/g, accounting for 5.6% to 76.3% of the total acid content.The malic acid content also showed similar ranges (0.40 to 19.26 mg/g) and percentages (from 12.9% to 79.6%) across the germplasms.No citric acid was detected in No.36, No.157, and No.203, while the highest citric acid content was found in No.64.PCA was performed to assess the impact of sugars and acids on the evaluation of peach germplasms (Figure 9).The first principal component (PC1) accounted for 49.5% of the variance, with sugars and acids providing the largest contributions.The second principal component (PC2) contributed 25.9% to the variance, with the ratio of sugars to acids and sucrose contributing significantly.Quinic, malic, and citric acids had negative loadings for PC2.Thus, PC2 reflects a taste profile emphasizing sweetness.No.107 was characterized by sweetness with light sourness, but No.155 was sour with a hint of sweetness.No.49 had negative loadings for both PC1 and PC2.Additionally, a close relationship was found between sweetness and total sugars, similar to the relationship between malic acids PCA was performed to assess the impact of sugars and acids on the evaluation of peach germplasms (Figure 9).The first principal component (PC1) accounted for 49.5% of the variance, with sugars and acids providing the largest contributions.The second principal component (PC2) contributed 25.9% to the variance, with the ratio of sugars to acids and sucrose contributing significantly.Quinic, malic, and citric acids had negative loadings for PC2.Thus, PC2 reflects a taste profile emphasizing sweetness.No.107 was characterized by sweetness with light sourness, but No.155 was sour with a hint of sweetness.No.49 had negative loadings for both PC1 and PC2.Additionally, a close relationship was found between sweetness and total sugars, similar to the relationship between malic acids and total acids.

Comparison of Instrumental and Sensory Results
A negative correlation existed between sweetness and quinic, malic, and citric acids (r > −0.7), while a positive correlation was observed between sweetness and the ratio of sugar to acids (r < 0.6) (Figure 10).Sourness negatively correlated with sweetness and the ratio of sugar to acids but positively correlated with acids.Persistence showed a positive correlation with acids and sourness and a negative correlation with sweetness and the sugar-to-acid ratio.Overall impression correlated negatively with acids and sourness while correlating positively with sweetness and the sugar-to-acid ratio.

Comparison of Instrumental and Sensory Results
A negative correlation existed between sweetness and quinic, malic, and citric acids (r > −0.7), while a positive correlation was observed between sweetness and the ratio of sugar to acids (r < 0.6) (Figure 10).Sourness negatively correlated with sweetness and the ratio of sugar to acids but positively correlated with acids.Persistence showed a positive correlation with acids and sourness and a negative correlation with sweetness and the sugar-to-acid ratio.Overall impression correlated negatively with acids and sourness while correlating positively with sweetness and the sugar-to-acid ratio.

Comparison of Instrumental and Sensory Results
A negative correlation existed between sweetness and quinic, malic, and citric acids (r > −0.7), while a positive correlation was observed between sweetness and the ratio of sugar to acids (r < 0.6) (Figure 10).Sourness negatively correlated with sweetness and the ratio of sugar to acids but positively correlated with acids.Persistence showed a positive correlation with acids and sourness and a negative correlation with sweetness and the sugar-to-acid ratio.Overall impression correlated negatively with acids and sourness while correlating positively with sweetness and the sugar-to-acid ratio.
Figure 10.Correlation matrix heat map.Heat map of spearman correlation between sensory and HPLC analysis between each biomarker for the overall population analysed.* Indicated significant differences between instrumental and sensory parameters according to Pearson correlation coefficient (p < 0.05).

Discussion
Fruit quality evaluation is an important component of peach fruit quality research.Sensory evaluation is commonly used to assess fruit quality.Based on the evaluation criteria, different samples are scored, allowing for the calculation of the mean values [20,21].This study included the sensory determination of 212 peach germplasm resources, focusing on aroma, palate (sweetness, sourness, astringency, and bitterness), persistence, and overall impression.In 212 peach germplasm resources, the predominant taste profiles, ranked from highest to lowest, were sweetness, sourness, astringency, and bitterness (Figure 1A).No.134 exhibited the most pronounced sweetness, with an excellent impression, resulting from its high total sugar content and low total acids content.The long-finish germplasms (No.38 and No.47) presented high total acid content (over 15 mg/g) and a small ratio of sugar to acid (below 5).The PCA revealed the influence of each sensory descriptor on the evaluation of peach germplasm resources.PC1 reflected an enjoyable taste, while PC2 corresponded with negative feedback.The sensory evaluation highlighted that No.211 was sweet with a great impression, but No.101 was astringent and bitter, resulting in an unpleasant impression.The profile of No.119 included light sweetness, strong sourness, and a long finish, leading to a harsh impression.These findings reveal the importance of a strong characteristic aroma in shaping consumer preferences.However, some germplasms can present a faulty flavour because of an excessively strong aroma.An overly strong passion fruit flavour had the potential to present bromhidrosis in one sensory evaluation [22].Thus, it was crucial to screen for peach germplasms with intense sweetness and an appropriate characteristic aroma to enhance peach quality.
An electronic nose (E-nose) is a simple, efficient, non-destructive, and stable device that imitates the human olfactory system.It has been used for assessing fruit quality and ripeness and in other applications [23,24].In our research, W5S, W1S, and W1W could explain up to 99.99% of the total contributions, indicating that these non-aromatic substance classes (nitrogen oxides, methane, and hydrogen sulphide) distinguished the peach fruit aromas in our research.The distinction in aromatic substance classes corresponded mainly to the response values of two sensors (W2W and W1C); these values aligned with studies on pear and blueberry quality and aroma evaluation [25,26].Upon comparing the results of the sensory evaluation and E-nose, the germplasms in Class I responded significantly to sensor W1C, which correlated with sensory perceptions of milk and caramel flavours.Germplasms in Class II showed a significant response to sensor W2W, associated with floral and herbaceous sensations.Notably, No.82, No.91, and No.193 were the most representative samples, with distinctive aromas of apple, butter/milk, and violet, respectively.These could serve as valuable indicators for fruit quality assessment and determining consumer preferences [27,28].Thus, they could be used in developing new peach germplasm with excellent aroma profiles.
Sucrose and malic acid were identified as the predominant components in peach germplasms, corroborating the results of previous studies [29].The correlation between the sensory and HPLC results indicated that the fruits' sweetness was primarily determined by their sugar content.The relative sweetness perceptions of fructose, sucrose, and glucose has been rated as 1.75, 1, and 0.75, respectively [16].Despite high levels of sugars, peach sweetness was not necessarily pronounced because of sucrose.Sucrose made up the largest share of total sugars and did not contribute as strongly to sweetness.Additionally, the types of sugars and their ratio with respect to acids also affected perceived sweetness.Malic acid, which imparted a soft, refreshing, slightly irritating, and long-lasting sour-bitter taste, was the primary acid in peaches, while citric acid provided a pleasant, quickly dissipating sour taste.Three germplasms (No.36, No.157, and No.203) lacked citric acid, resulting in less sourness detected in their sensory evaluations.Furthermore, for the 212 peach germplasms studied, it was observed that a sugar-to-acid ratio exceeding 20 corresponded to sweetness scores above 5.This finding supports the results of the research by Colaric et al. [14], who suggested that a sugar-to-acid ratio ranging from 20 to 60 is typically associated with a moderate level of perceived sweetness.This ratio is important because it provides a balance of flavours, contributing to the overall sensory pleasure elicited by and the acceptability of peach fruit by consumers.By analysing the sugar-to-acid ratio, breeders and researchers can better understand and predict the sweetness profiles of peaches, constituting a crucial factor in consumer satisfaction and the marketability of the fruit.

Conclusions
In this study, we employed sensory evaluation, an electronic nose (E-nose), and highperformance liquid chromatography (HPLC) to assess fruit quality attributes such as aroma, sugars, and acids across 212 peach germplasm resources.These methods significantly contributed to the preliminary screening of peach germplasms with specific desirable characteristics.This study provides essential information that could facilitate the creation of superior germplasm resources.Therefore, evaluating and analysing the aroma and taste profiles of peach germplasms hold significant theoretical and practical value for the discovery, development, and utilization of high-quality peach resources.For future research, the integration of the electronic nose technique with headspace solid-phase microextraction (HS-SPME) and gas chromatography-mass spectrometry (GC-MS) could enhance the discrimination of peach fruit aroma components and aid in identifying the key aroma compounds within peach germplasm resources.This multidisciplinary approach may lead to a deeper understanding of the complex interactions between volatile compounds and contribute to the advancement of peach breeding programs targeting improved fruit aromas and flavour.

Figure 3 .
Figure 3. Hierarchical cluster analysis of E-nose performance for 212 peach germplasm resourc red, Class I germplasms; blue, Class II germplasms; green, Class III germplasms.

Figures 6 Figure 5 .
Figures 6 and 7 show the stable signals for Classes I and II in the response values of the three primary sensors and the three aromatic sensors.The response values of the three main sensors differed significantly between Class I and Class II.Compared with the response values of the three aroma sensors, germplasms in Class I responded notably to the sensor W1C, while germplasms in Class II responded significantly to sensor W2W.This result was consistent with the previous results for the coefficient of variation.It indicates that the differences in peach aromas predominantly arose from non-aromatic substances

Figures 6
Figures 6 and 7 show the stable signals for Classes I and II in the response values of the three primary sensors and the three aromatic sensors.The response values of the three main sensors differed significantly between Class I and Class II.Compared with the

Figure 5 .
Figure 5. No.193 response values to different sensors of an electronic nose.

Figures 6
Figures6 and 7show the stable signals for Classes I and II in the response values of the three primary sensors and the three aromatic sensors.The response values of the three main sensors differed significantly between Class I and Class II.Compared with the response values of the three aroma sensors, germplasms in Class I responded notably to the sensor W1C, while germplasms in Class II responded significantly to sensor W2W.This result was consistent with the previous results for the coefficient of variation.It indicates that the differences in peach aromas predominantly arose from non-aromatic substances such as methane, nitrogen oxides, and hydrogen sulphide, with variations in aromatic substances primarily reflected in the response values of W2W and W1C.

Figure 10 .
Figure10.Correlation matrix heat map.Heat map of spearman correlation between sensory and HPLC analysis between each biomarker for the overall population analysed.* Indicated significant differences between instrumental and sensory parameters according to Pearson correlation coefficient (p < 0.05).

Figure 10 .
Figure10.Correlation matrix heat map.Heat map of spearman correlation between sensory and HPLC analysis between each biomarker for the overall population analysed.* Indicated significant differences between instrumental and sensory parameters according to Pearson correlation coefficient (p < 0.05).

Table 2 .
Special flavours in 212 peach germplasm resources.
presents the sucrose, glucose, fructose, and sorbitol content of 212 germplasms determined via HPLC, with sucrose ranging from constituting 11.9% to Foods 2023, 12, x FOR PEER REVIEW 11 of 19 83.3% (from 4.61 mg/g to 118.28 mg/g) of the total sugar content, while sorbitol corresponded to the lowest content, ranging from 0.01 to 26.80 mg/g (from 0% to 18.6%), in the 212 germplasms.The glucose and fructose percentages were similar, spanning from 3.8% to 47.7%.No.158 had the lowest total sugar content at 25.86 mg/g, whereas the highest content was 172.61 mg/g, corresponding to No.73.In Figure 8B, the total acid content varies from 2.07 mg/g in No.150 to 33.61 mg/g in No.76.The quinic acid content ranges from 0.32 to 10.72 mg/g, accounting for 5.6% to 76.3% of the total acid content.The malic acid content also showed similar ranges (0.40 to 19.26 mg/g) and percentages (from 12.9% to 79.6%) across the germplasms.No citric acid was detected in No.36, No.157, and No.203, while the highest citric acid content was found in No.64.germplasms: blue area corresponds to quinic acid; orange area corresponds to malic acid; and grey area corresponds to citric acid.