Characterization of aroma-active volatiles in three Chinese bayberry (Myrica rubra) cultivars using GC–MS–olfactometry and an electronic nose combined with principal component analysis

    Characterization of Aroma-active Volatiles in Three Chinese Bayberry (Myrica rubra) Cultivars Using GC–MS–Olfactometry and an Electronic Nose Combined with Principal Component Analysis Huan Cheng, Jianle Chen, Shiguo Chen, Dan Wu, Donghong Liu, Xingqian Ye PII: DOI: Reference:

S0963-9969(15)00088-5 doi: 10.1016/j.foodres.2015.03.006 FRIN 5711

To appear in:

Food Research International

Received date: Revised date: Accepted date:

25 July 2014 21 February 2015 1 March 2015

Please cite this article as: Cheng, H., Chen, J., Chen, S., Wu, D., Liu, D. & Ye, X., Characterization of Aroma-active Volatiles in Three Chinese Bayberry (Myrica rubra) Cultivars Using GC–MS–Olfactometry and an Electronic Nose Combined with Principal Component Analysis, Food Research International (2015), doi: 10.1016/j.foodres.2015.03.006

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Characterization of Aroma-active Volatiles in Three

T

Chinese Bayberry (Myrica rubra) Cultivars Using

IP

GC–MS–Olfactometry and an Electronic Nose Combined

SC R

with Principal Component Analysis Huan Cheng, Jianle Chen, Shiguo Chen, Dan Wu, Donghong Liu, and Xingqian Ye*

NU

Zhejiang University, College of Biosystems Engineering and Food Science; Fuli Institute of Food Science; Zhejiang Key Laboratory for Agro-Food Processing; Zhejiang R & D Center for Food

MA

Technology and Equipment. Hangzhou 310058, China

ABSTRACT: Chinese bayberry (Myrica rubra Sieb. et Zucc.) is one of the most

D

popular and valuable fruits in China because of its unique and exquisite flavor. In this

TE

study, headspace solid-phase micro-extraction (HS-SPME) coupled with gas

CE P

chromatography–mass spectrometry and olfactometry (GC–MS–O) analyses, were used to characterize the aroma-active profiles of the fruits from three different bayberry cultivars. The aim was to differentiate the bayberry cultivars by their aroma.

AC

Fifty-five volatile components, composed of aldehydes (10), alcohols (9), esters (8), terpenes (17) and others (11), were identified by optimized HS-SPME/GC–MS. Meanwhile, 36 aroma-active compounds were detected by olfactometry using detection frequency analysis (DFA). Hexanal (grass-like), (E)-2-hexenal (green), nonanal (fruit, flower), 1-hexanol (flower), and isocaryophillene (wood) were identified in all three cultivars. Further principal component analysis (PCA) of the active aromas revealed their contributions to the odor differences among the bayberry cultivar groups. The BQ bayberry was characterized by having a stronger “herb” odor, which is mainly caused by benzoic acid and methyl ester. DK bayberry had a stronger “grass” odor, which is mainly caused by 2,6-dimethyl-2,4,6-octatriene, while FHZ bayberry had a stronger “pine” odor, which is caused mainly by α-pinene. The GC–MS–O and electronic nose techniques, when combined with PCA, could be used

ACCEPTED MANUSCRIPT to successfully distinguish between different bayberry cultivars.

Keywords: Bayberry (Myrica rubra); gas chromatography-mass spectrometry (GC-MS-O);

principal

component

analysis

(PCA);

E-nose;

T

-olfactometry

IP

aroma-active.

SC R

1. Introduction

Chinese bayberry (Myrica rubra Sieb. et Zucc) fruits are one of the

NU

most appealing fruits in the markets because of their unique flavor, taste,

MA

and high levels of nutritional components such as carbohydrates, sugars, organic acids, minerals, vitamins, and polyphenols (Cheng et al., 2009;

D

Fang et al., 2009; Zhou et al., 2009). However, they are quite perishable

TE

and have short fruiting seasons. Therefore, in order to increase their

CE P

consumption and their shelf life, the flesh from bayberry fruits is processed into sweets, jam, juice, and wine, or canned in syrup (Shao &

AC

He, 2007). Bayberry fruit juice and wine are also important Chinese export products, which means that bayberry has a relatively high export value (Fang & Bhandari, 2012). Aroma is one of the most valued attributes of Chinese bayberry fruits, and can greatly influence the consumers’ acceptance of the bayberry fruit and related fruit-products. Gas chromatography-mass spectrometry (GC-MS) has been widely used to analyze the aroma composition of fruits. However, not all volatile compounds are aroma-active in some products due to different odor thresholds and interactions between

ACCEPTED MANUSCRIPT compounds. One of the major difficulties when studying odors is the identification of those compounds that really contribute to the food flavor.

IP

T

A combination of GC–MS with olfactometry and detection frequency

SC R

analysis (DFA) has been used to identify odor-active compounds (Pang, Chen, Hu, Zhang & Wu, 2012; Pang, Guo, Qin, Yao, Hu & Wu, 2012). It has been extensively used in the flavor analysis of a number of fruits,

NU

especially oranges (Plotto, Margaria, Goodner & Baldwin, 2008),

MA

cherry(Sun, Jiang & Zhao, 2010), muskmelon(Pang et al., 2012b), guava (Pino & Bent, 2013), papaya(Pino, 2014), and banana(Pino & Febles,

D

2013). However, there is limited information available on the volatile

TE

compositions of fruits from different bayberry varieties and cultivars, and

CE P

very few studies have focused on the aroma-active components of Chinese bayberry. Kang et al. (2012) reported that ethyl acetate,

AC

1-hexanol, (Z)-3-nonen-1-ol, and β-caryophyllene were the predominant flavor volatiles. However, studies on the aroma-active volatiles, instead of volatiles in general, are more important when it comes to the development of Chinese bayberry products. Chinese bayberry cultivars and plants that have been grown in different locations have different flavors, which hinders product development and reduces economic value. Furthermore, it is quite difficult to differentiate between the flavors of different Chinese bayberry cultivars, especially when it has been processed into a product. To our best knowledge, there

ACCEPTED MANUSCRIPT have been no studies that have classified the volatile profiles and the aroma-active compounds produced by different Chinese bayberry

IP

T

(Myrica rubra Sieb. et Zucc.) cultivars using GC–MS–O, with or without

SC R

PCA analysis, which is a multivariate statistics method.

GC–MS studies have mainly focused on the measurement of certain volatile components, while the electronic nose (e-nose) is an instrument

NU

that uses chemical sensors to detect volatiles and then provide a pattern

MA

output showing the component combinations contributing to a defined smell. It has been successfully used to classify olive oil (Haddi et al.,

D

2013), differentiate between ginseng made from varieties that have

TE

different varietal origins (Li et al., 2012), identify different milk

CE P

flavorings (Wang, Xu & Sun, 2010), authenticate different cherry tomato juices (Xuezhen, Jun & Shanshan, 2014) and characterize different

AC

famous liquors (Xiao, Yu & Niu, 2014). Thus, it could potentially be used to classify different bayberry samples. The aim of this study was to detect differences in the volatile composition of three bayberry cultivars. These differences could then be used to identify the different bayberry cultivars. Principal component analysis (PCA) grouped the samples according to their volatile profile. The data from the GC–MS–O analysis and the electronic nose were then statistically analyzed in order to detect the volatile compounds that could be used to differentiate the bayberry cultivars according to their cultivar

ACCEPTED MANUSCRIPT group. 2. Materials and methods

IP

T

2.1. Fruit materials

SC R

Bayberry cultivars that produced fruits with good edible qualities and had been planted over large areas of land were collected in Zhejiang Province, China, during May and June, 2013. The following cultivars

NU

were collected: BQCX (Biqi, CiXi), BQXJ (Biqi, XianJu), DKNH

MA

(Dongkui, NingHai), DKXJ (Dongkui, XianJu), and FHZ (Fenhongzhong, ShangYu). This meant that there were three cultivar groups: BQ, DK, and

D

FHZ. After the fruit had been picked, they were kept cold in baskets and

TE

immediately transported to the laboratory where they were frozen in

2.2.

CE P

liquid nitrogen within 12 h and stored at –80°C until needed for analysis. Chemicals

AC

A mixture of n-alkanes (C8–C20) was used for the retention index (RI) analyses. The experimental procedure and the RI calculation were carried out according to the chemical manufacturer’s instructions (Sigma Chemical Co., St. Louis, MO, USA). The cyclohexanone (9.46 mg/kg of fruit juice), used as the internal standard, was purchased from J&K Chemical Ltd (Shanghai, China). The sodium chloride used for volatile extraction and the other reagents were all either analytical grade or the highest purity that was commercially available. Four

different

coating

fibers

for

Headspace

Solid-phase

ACCEPTED MANUSCRIPT Micro-extraction (HS-SPME) were tested. These were 100 µm polydimethylsiloxane (PDMS), 75 µm carboxen/polydimethylsiloxane µm

polydimethylsiloxane/divinylbenzene

T

65

IP

(CAR/PDMS),

SC R

(PDMS/DVB), and 50/30 µm DVB/CAR/PDMS. They were purchased from Supelco, Inc. (Bellefonte, PA, USA).

2.3. Optimal conditions for headspace sampling and SPME

NU

A manual SPME (Supelco, Inc.) fiber was used for volatile extraction

MA

after the fiber had been conditioned. The flesh was ground to pulp by a commercial blender. Then the bayberry pulp (4 g) and the internal

D

standard (10 µL)were immediately transferred to a 15 mL headspace

TE

bottle containing 4 g of sodium chloride (NaCl) saturated solution (Aprea

CE P

et al., 2012) The mix was then equilibrated with a laboratory stirrer/hot plate (model PC-420, Corning Inc. Life Science, Acton, MA, USA). After

AC

the solution had equilibrated, a stainless steel needle, housing the SPME fiber, was placed through a hole in the top of the headspace bottle and fed through until it was 1 cm above the liquid surface. Then the samples were magnetically stirred at 800 rpm. Three independent extractions were carried out for each bayberry sample. In order to improve volatile compound absorption, the following experimental parameters were investigated: four coating fibers (PDMS, CAR/PDMS, PDMS/DVB and DVB/CAR/PDMS); incubation times between 5 and 20 min; extraction temperatures between 35°C and 65°C;

ACCEPTED MANUSCRIPT and extraction times between 10 and 40 min. The analysis was conducted in triplicate for each parameter investigated.

T

GC–MS analysis

IP

2.4.

SC R

The SPME extract was injected into the port of an Agilent 7890A-5975C GC/MSD equipped with a DB-5 capillary column (30 m × 0.25 mm, 0.25 µm film thickness) (Agilent Technologies) and desorbed

NU

at 250°C for 3 min. The injection port was operated in split mode (1:10),

MA

and 99.999% pure helium was used for vial pressurization and as the carrier gas. The helium flow rate was 1.4 mL/min. The initial oven

D

temperature was 40°C (2 min), which was ramped up at 5°C/min to

TE

170°C and held there for 2 min. Then it was ramped up at 10°C/min to

CE P

250°C and held there for 5 min. The mass detector was operated in the electronic impact (EI) mode at 70 eV and the source temperature was set

AC

at 250°C. The mass spectra were scanned in the m/z 29–350 amu range at 1 s intervals. 2.5.

Detection frequency analysis by GC–MS–O

The odor-active compounds were characterized by a sniffing port (Sniffer 9000, Brechbühler, Switzerland) coupled to a GC–MS (7890A-5875C, Agilent Technologies, Inc.). At the exit of the capillary column, the effluents were split 1:1 (by volume) into a sniffing port and an MS detector by employing Agilent capillary flow technology. The transfer line to the GC–O sniffing port was held at 260°C. The GC–MS

ACCEPTED MANUSCRIPT conditions were the same as those described above. An eight-membered panel of assessors was required to individually

IP

T

sniff the GC effluent and report their results by the GC–O analyses. The

SC R

panel consisted of a mixed group of both sexes, aged between 20 and 30 years. The panelists were trained prior to the sensory analysis so that they could become familiar with the odor descriptions for solutions of artificial

NU

odorants and different bayberry samples. In total, eight GC –O runs were

MA

performed (one run for each assessor). The aroma-active compounds perceived by the panelists were recorded when sniffing the effluent from

D

the sniffing mask. The panelists also noted the perceived odor

TE

characteristics and intensities. At the sniffing port, any odorant that had

CE P

total detection frequencies ≥ 2 was arbitrarily considered to have potential aroma activity (Pang et al., 2012a). Electronic nose system

AC

2.6.

An electronic nose (e-nose, PEN2, Airsense Analytics, GmBH, Schwerin, Germany) was used to discriminate between the odor patterns produced by the different cultivars. The e-nose consists of a sampling apparatus that is exposed to the volatiles, an array of sensors composed of ten different metal-oxide semiconductors (MOS) enclosed in a small chamber, and computer-controlled pattern recognition software. Each sample was placed in an airtight 500 mL glass jar (concentration chamber). The glass jar was then closed and the headspace inside was

ACCEPTED MANUSCRIPT equilibrated for 30 min. The measurement process consisted of three different phases: concentration, measurement, and stand-by. Airflow was

IP

T

always kept constant throughout the concentration chamber during the

SC R

three phases. The measurement phase lasted 70 s and the collected data interval was 1 s so that the sensors could reach a stable value. When a measurement was completed, a stand-by phase was activated (70 s) in

NU

order to clean the circuit and return the sensors to their baseline. All the

MA

e-nose measurement procedures were carried out at 25 ± 1°C. Each analysis was repeated more than four times, and all of the sensor response

Statistical data analysis

TE

2.7.

D

data were analyzed by the e-nose software.

CE P

Significant volatile constituent differences among the bayberry fruits were determined by one-way analysis of variance (ANOVA) using SPSS,

AC

version 17.0 (SPSS Inc., 2009). PCA of volatile compounds was performed using Unscrambler v. 9.7 (CAMO AS, Trondheim, Norway) software. The aims of the PCA were to reduce the number of variables and to remove redundant information. The data were normalized by setting the mean values to zero and scaling on the basis of one standard deviation. The first principal component (PC1) was the axis, which contained the largest possible amount of information,

and

the

perpendicular to PC1.

second

principal

component

(PC2)

was

ACCEPTED MANUSCRIPT 3. Results and discussion 3.1. Optimization of SPME conditions

IP

T

The aroma compounds in Chinese bayberry fruits were extracted using

SC R

HS-SPME and the highest peak area response was selected in order to optimize the main parameters. Different extraction conditions were tested. These were four different coating fibers, incubation time, extraction

NU

temperature, and extraction time, and were investigated and optimized

MA

based on the total ion response in the GC–MS, as described previously (Lin, Zhuang, Lei, Yang & Zhao, 2013).

D

Fig.1.a showed the effect of four different coating fibers (PDMS,

TE

CAR/PDMS, PDMS/DVB and DVB/CAR/PDMS) on the adsorption of

CE P

volatile compounds from the bayberry samples. The results shows that there were significant differences among the four coating fibers (p<0.05) ,

AC

and the DVB/CAR/PDMS fiber obtained the largest peak area for the volatiles. Therefore, the DVB/CAR/PDMS fiber was selected for the experiments.

Fig.1.b shows the effect of incubation time (5, 10, 15, and 20 min) on the detection of total volatiles. The total volatile amounts significantly rose with increasing incubation time. However, there was no significant difference between 15 min and 20 min (p < 0.05), which indicated that a 15 min incubation time would allow distributions between the fiber, the vial headspace, and the analytes to reach an equilibrium. Similarly, Figs.

ACCEPTED MANUSCRIPT 1.c and d show that the optimum extraction time was 30 min and the optimum temperature was 55°C, respectively.

IP

T

The optimal extraction conditions were as follows: DVB/CAR/PDMS;

SC R

incubation time, 15 min; extraction temperature, 55°C; and extraction time, 30 min. These conditions were applied during the extraction of volatile compounds from bayberry fruits.

NU

3.2. SPME/GC–MS analysis of volatile composition characteristics

MA

After SPME, using the optimized conditions, the volatile compounds were separated on a DB-5 column and identified by GC–MS. The

D

detected compounds and their retention indices (RI) are shown in Table 1.

TE

Aldehydes (10), alcohols (9), esters (8), and terpenes (17) were the most

CE P

abundant compounds in bayberry fruit among the 55 volatile components identified and quantified by comparing the mass spectra. The results were

AC

confirmed by comparing the RI with data from the literature (Kang, Li, Xu, Jiang & Tao, 2012; Xu, Zhang, Fang, Sun & Wang, 2014). Terpenes were the most dominant volatiles in bayberry fruits and could be described as typical bayberry flavor constituents, which was similar to the results reported by Kang et al. (2012) and Xu et al. (2014). Aldehydes were listed as the second largest volatile group, and included hexanal (A1), (E)-2-hexenal (A2), nonanal (A8), and decanal (A10), which were present in all five bayberry cultivars. Several volatile compounds were reported for the first time in Chinese bayberry, e.g. (E)-2-hexenal (A2)

ACCEPTED MANUSCRIPT and decanal (A10) (Table 1). The content of esters were relatively lower than aldehydes and terpenes

IP

T

in bayberry fruit (Table 1). It was different from previously report by Xu

SC R

et al. that the esters were listed as the major volatile, and ethyl acetate was classified as the top-note flavor of bayberry juices (Xu et al., 2014). These differences may be caused by the different cultivars, different

NU

detector and storage time.

MA

Alcohols were also present at high concentrations in the bayberry cultivars. They may be derived from the oxidative degradation of fatty

D

acids as previously described (Lin et al., 2013). Among the alcohols, B2

TE

(1-hexanol) and B3 (3,7-dimethyl-1,6-octadien-3-ol) were detected in all

CE P

the bayberry cultivars, and B3 was the most abundant alcohol detected in the DKNH samples.

AC

A comparison of the three Chinese bayberry cultivars (Table 1) showed that the terpenes group contained the largest number of compounds. The total profile of the volatile constituents of all cultivars were predominate by β-caryophyllene (D11) except FHZ, which was most abundant with α-pinene (D1) (42.46µg/g). The concentration of β-caryophyllene (D11) in Chinese bayberry fruit was much higher than other fruits, such as cherry (Wen et al., 2014), passion fruit (Janzantti, Macoris, Garruti & Monteiro, 2012), papaya (Pino, 2014), pomegranate (Caleb, Opara, Mahajan, Manley, Mokwena & Tredoux, 2013), kiwifruit (Garcia, Quek,

ACCEPTED MANUSCRIPT Stevenson & Winz, 2012) and strawberry (Samykanno, Pang & Marriott, 2013) , which may indicated that β-caryophyllene is one of the important

IP

T

aroma-active compounds.

SC R

In summary, the BQ and DK cultivar groups contained larger numbers of terpenes, such as β-caryophyllene (D11) and caryophyllene oxide (E9),

camphene (D2), and β-pinene (D3).

NU

than FHZ and the FHZ cultivar contained higher levels α-pinene (D1),

MA

3.3. Analysis of aroma-active volatiles by GC–MS–O Although the aroma compositions of different cultivars were identified

D

by GC–MS, it was still not clear which of the aromas truly contributed to

TE

the flavor of the Chinese bayberry. Therefore, GC–MS combined with

CE P

olfactometric analysis was used to identify the aroma-active compounds in the different Chinese bayberry cultivars. The procedure was similar to

AC

the one used in our previous study (Cheng, Qin, Guo, Hu & Wu, 2013). The results of the olfactometric analysis are summarized in Table 2. Among the 56 compounds that were identified in the different Chinese bayberry cultivars by GC–MS–O, 36 of them were identified as aroma-active compounds by the DFA method (detection frequency ≥ 2), and 31 were identified using RI, GC–MS, and odor properties (Table 2). The identified aroma-active substances included nine aldehydes, six alcohols, four esters, six terpenes, six others, and five unknown compounds that had concentrations below the GC–MS detection

ACCEPTED MANUSCRIPT threshold. Two unknown compounds (RI = 972 and 1020) were tentatively identified as hexanoic acid and α-phellandrene, respectively,

IP

T

according to their RI, odor description and/or other relevant references

SC R

(Kang et al., 2012). However, further studies are needed to definitively identify these unknown components using more effective extraction methods, more sensitive detection techniques, and different standards.

NU

The analysis by DFA revealed that the aroma-active compounds

MA

identified in bayberry fruit covered a wide diversity of structures and chemical functions that can be classified into five classes described as

TE

compounds.

D

“grassy”, “mushroom”, “fruity, floral”, “cucumber” and “spicy-anise”

CE P

The DFA revealed that the aroma-active compounds identified in bayberry fruit covered a wide range of structures and chemical functions

AC

that could be classified into five classes. These were “grassy”, “mushroom”, “fruity, floral”, “cucumber”, and “spicy-anise” compounds. The total number of aroma-active compounds varied between the cultivars: The two cultivars in the BQ group contained 27 compounds, in which BQCX and BQXJ contained 21 and 19 compounds, respectively (Tables 1 and 2). Among these 27 aroma-active compounds, four compounds: A1, A8, C7, and E2, were identified as being the most important aroma-active compounds (DF ≥ 6). Further aroma descriptions were based on an evaluation of their odor, and it was concluded that A1

ACCEPTED MANUSCRIPT may contribute to the “grassy” odor (Table 2); and the “mushroom” odor might be caused by the presence of D6, E3, and E4, which were only

IP

T

present in the BQCX samples. The fruity, floral odor might be caused by

SC R

A2, A6, A8, B2, and E2; the “cucumber” odor might be caused by A9 and B4; and the “spicy-anise” odor might be due to the presence of D11. Furthermore, C7 (3-nonenoic acid, methyl ester), which contributes

NU

towards the “oil, fat” odor, was found in both BQ cultivars; and C1 (ethyl

MA

acetate), which contributes towards the “pineapple” odor, was found in the BQCX cultivar. These results confirmed the results of a previous by

Kang

et

al.

(2012).

In

contrast,

E8

D

study

TE

(dihydro-5-pentyl-2(3H)-furanone), which contributes towards the “spicy”

CE P

odor, was only found in the BQXJ cultivar. In the DK bayberry cultivars (two samples), a total of 23 aroma-active

AC

compounds (14 for the DKXJ cultivar and 20 for the DKNH cultivar, respectively) were identified. Among them, three compounds (A1, A8, and D6) were the most important aroma-active compounds (DF ≥ 6) (Tables 1 and 2). According to the odor evaluation, the “grassy” odor was caused by A1 and E6 in DKXJ, and by A1 in DKNH; The “mushroom” odor might be due to the presence of D6 and E4; the “fruity, floral” flavor might be caused by A2, A6, A8, B2, B3, C1, and C2 in DKXJ, and by A2, A8, B2, B3, and C1 in DKNH. B3 (linalool), which was described as being a plum-like, fragrant aroma in a previous report (Lin et al., 2013),

ACCEPTED MANUSCRIPT also made a significant contribution to the characteristic aroma produced by DK. The “cucumber” flavor might be caused by A9 in DKXJ and by

IP

T

B4 in DKNH; and the “spicy-anise” odor may be due to the presence of

SC R

D11 in the DKNH cultivar.

A total of 18 aroma-active compounds were identified in the FHZ cultivar (Tables 1 and 2). Five compounds (A1, A8, B2, D1, and D3)

NU

were the most important aroma-active compounds among the 18 found in

MA

the FHZ cultivar (DF ≥ 6). Based on the odor evaluation, it is probable that the “grassy” odor produced by FHZ fruit may have been caused by

D

A1; the “mushroom” odor was due to the presence of E4; the “fruity,

TE

floral” odor was possibly caused by A2, A8, B2, B3, and D4; and the

CE P

“cucumber” odor may have been caused by B4 and E7, which was only found in the FHZ samples. The “spicy-anise” odor might be due to the

AC

presence of B6 and D11. Furthermore, a unique “pine, turpentine” odor, caused by D1 (α-pinene) and D3 (β-pinene), and a “lemon, citrus” odor, caused by D4 (D-limonene), were also detected in the FHZ fruit samples. Five components, namely hexanal (A1), (E)-2-hexenal (A2), nonanal (A8), 1-hexanol (B2), and isocaryophillene (D10), were detected by the panelists in all the bayberry cultivars, which indicated that they contributed actively to the aroma profile of bayberry fruit. Interestingly, hexanal (A1) had been reported as being active in an earlier study on bayberry odors (Kang et al., 2012).

ACCEPTED MANUSCRIPT 3.4. Principal component analysis (PCA) of GC-MS-O data Principal component analysis (PCA), multivariate pattern recognition

IP

T

procedure, was applied to determine if there were differences in the

SC R

aroma-active volatile patterns from the five bayberry fruit (Lignou, Parker, Oruna-Concha & Mottram, 2013). In order to improve interpretation of the obtained data (Table 2), the 31 GC–MS–O signals

NU

produced by the active volatiles in the bayberry samples were manually

MA

integrated and their relative percentages were analyzed by PCA. The bayberry cultivars were successfully divided into the three cultivar groups

D

(BQ, DK, and FHZ) on the basis of the relationships between the

TE

cultivars (scores) and their active volatile compound contents (loadings)

CE P

(Fig. 2).

In Fig. 2.a, the scores scatter plot for the two first principal components

AC

(PC1 and PC2) represent the differences among the bayberry samples. Fig. 2.b represents the corresponding loadings plot, which established the relative importance of each active volatile compound, and the relationships between the volatile compounds and the samples. Generally, when the PCs have more than 85% cumulated reliability, compared to the original dataset, then these PCs can be used to replace the original data (Cheng et al., 2013). The first two principal components (PC1 and PC2) accounted for 98% and 2% of the total variation, respectively, which separated FHZ from the other cultivars (Fig. 2).

ACCEPTED MANUSCRIPT The BQ bayberry cultivars (BQCX and BQXJ) were located in the negative region of PC2, which was clearly isolated from the other

IP

T

bayberry samples. The A9, C5, E3, and E8 compounds were associated

SC R

with the BQ cultivar group (Fig. 2.b and Table 2). This indicated that (E)-2-nonenal (A9), benzoic acid, methyl ester (C5), 1-decen-3-one (E3), and dihydro-5-pentyl-2(3H)-furanone (E8) were important when it came

NU

to differentiating the two BQ cultivars from the other cultivars.

MA

The DK bayberry cultivars (DKNH and DKXJ) were located in the positive regions of PC1 and PC2, which related to the B3, B9, D11, D14,

D

E6, E10, and E11 volatile compounds. This suggested that they were

TE

important when it came to differentiating the DK cultivars from the other

CE P

bayberry cultivars (Fig. 2.b and Table 2). The FHZ bayberry cultivar was located in the negative region for PC1

AC

and the positive region for PC2, and was clearly differentiated from the other bayberry cultivars. FHZ was characterized by the variables associated with the positive values in the two first principal components, primarily, B7, D1, D3, and E7. The volatile profile for FHZ contained a higher proportion of α-terpineol (B7), α-pinene (D1), β-pinene (D3), and pinocarvone (E7) (Fig. 2.b and Table 2). Therefore, these volatiles play a major role in differentiating FHZ from the other bayberry cultivars. In this study, we attempted to classify and differentiate the bayberry samples according to their aromatic profile characteristics using the

ACCEPTED MANUSCRIPT SPME extraction technique coupled with GC–MS–O analysis and data analysis by PCA. However, our results may have wrongly classified some

IP

T

cultivars because the number of bayberry cultivars analyzed in this study

SC R

was relatively small for accurate generalization. In order to minimize the possibility of false results, future studies should increase the number of cultivars they analyze.

NU

3.5. Identifying different bayberry cultivars using e-nose

MA

We used the e-nose equipment to differentiate the bayberry cultivars in order to confirm the results of the GC–MS–O combined with the PCA

D

technique. The score plot for the first two principal components (PC1

TE

versus PC2) is shown in Fig. 3. The results indicated that bayberry

CE P

cultivars that have different botanical origins can be distinguished according to their odors using the e-nose combined with PCA. The score

AC

plot revealed that separation along PC1 accounted for 100% of the variation in the sample set. Each cultivar group was clearly distinguished from the other groups by the PCA analysis (Fig. 3), and there were obvious differences among the different cultivars. The bayberry samples were separated into the five different cultivars (BQCX, BQXJ, DKNH, DKXJ, FHZ). The score plot (Fig. 3.) also reveals that separation along PC1 and PC2 accounted for 100% of the variation in the sample. Therefore, according to the obtained results, the e-nose can be used as a

ACCEPTED MANUSCRIPT quick analysis tool that can be used to distinguish between samples from different cultivars by measuring the concentration of volatile compounds.

IP

T

The findings were in good agreement with the results obtained by

SC R

GC–MS–O as they both separated the samples into the different groups. This confirmed the usefulness of the e-nose as a bayberry fruit identification tool.

NU

4. Conclusions

MA

In this study, the volatile profiles of different bayberry cultivars were evaluated by combining the SPME/GC–MS–O and e-nose techniques

D

with multivariate statistical analysis. PCA revealed that there were

TE

significant differences among the bayberry cultivars, and that it was

CE P

possible to classify the samples into the cultivar groups using GC–MS–O and the e-nose. Hexanal (grass-like), (E)-2-hexenal (green), nonanal (fruit,

AC

flower), 1-hexanol (flower), and isocaryophillene (wood) were detected as aroma-contributing compounds that were common to all three cultivars. The odors of the key active aromas that most correlated with the differences between the bayberry cultivars were also identified in this research. The BQ cultivars were characterized as having a stronger “herb” odor; the DK group had a stronger “grass” odor, and the FHZ cultivar had a stronger “pine” odor. This combination of headspace analysis and chemometric methods can be used to successfully discriminate between different bayberry cultivars.

ACCEPTED MANUSCRIPT Acknowledgements This study was supported financially by the National Key Technology

IP

T

R&D Program (2012BAD31B06, 2012BAD31B02) and Zhejiang

SC R

Province Science and Technology Project (2012T2T123, 2010R50032). We thank Dr. Wu from the Experiment Education Centre for offering

NU

technical assistance and experimental facilities for our experiment.

MA

References: Aprea, E., Corollaro, M. L., Betta, E., Endrizzi, I., Demattè, M. L., Biasioli, F., & Gasperi, F. (2012). Sensory and instrumental profiling of 18 apple cultivars to investigate the relation between perceived

D

quality and odour and flavour. Food Research International, 49(2), 677-686.

TE

Caleb, O. J., Opara, U. L., Mahajan, P. V., Manley, M., Mokwena, L., & Tredoux, A. (2013). Effect of modified atmosphere packaging and storage temperature on volatile composition and postharvest life of minimally-processed pomegranate arils (cvs. 'Acco' and 'Herskawitz'). Postharvest Biology And

CE P

Technology, 79, 54-61.

Cheng Jiyu, Zhou Shaohuan, Wu Dan, Chen Jianchu, Liu Donghong, & Ye Xingqian (2009). Bayberry (Myrica rubra Sieb. et Zucc.) kernel: A new protein source. Food Chemistry, 112, 469-473. Cheng, H., Qin, Z. H., Guo, X. F., Hu, X. S., & Wu, J. H. (2013). Geographical origin identification of

AC

propolis using GC–MS and electronic nose combined with principal component analysis. Food Research International, 51(2), 813-822. Fang Zhongxiang, Zhang Yuhuan, Lü Yuan, Liu Donghong, & Ye Xingqian (2009). Phenolic compounds and antioxidant capacities of bayberry juices. Food Chemistry, 113, 884-888. Fang, Z., & Bhandari, B. (2012). Comparing the efficiency of protein and maltodextrin on spray drying of bayberry juice. Food Research International, 48(2), 478-483. Garcia, C. V., Quek, S. Y., Stevenson, R. J., & Winz, R. A. (2012). Characterisation of bound volatile compounds of a low flavour kiwifruit species: Actinidia eriantha. Food Chemistry, 134(2), 655-661. Gokbulut, I., & Karabulut, I. (2012). SPME-GC-MS detection of volatile compounds in apricot varieties. Food Chemistry, 132(2), 1098-1102. Haddi, Z., Alami, H., El Bari, N., Tounsi, M., Barhoumi, H., Maaref, A., Jaffrezic-Renault, N., & Bouchikhi, B. (2013). Electronic nose and tongue combination for improved classification of Moroccan virgin olive oil profiles. Food Research International, 54(2), 1488-1498. Janzantti, N. S., Macoris, M. S., Garruti, D. S., & Monteiro, M. (2012). Influence of the cultivation system in the aroma of the volatile compounds and total antioxidant activity of passion fruit. LWT-Food Science and Technology, 46(2), 511-518. Kang, W., Li, Y., Xu, Y., Jiang, W., & Tao, Y. (2012). Characterization of Aroma Compounds in

ACCEPTED MANUSCRIPT Chinese Bayberry (Myrica rubra Sieb. et Zucc.) by Gas Chromatography Mass Spectrometry (GC-MS) and Olfactometry (GC-O). Journal Of Food Science, 77(10), C1030-C1035. Lignou, S., Parker, J. K., Oruna-Concha, M. J., & Mottram, D. S. (2013). Flavour profiles of three novel acidic varieties of muskmelon (Cucumis melo L.). Food Chemistry, 139(1–4), 1152-1160.

T

Lin, L., Zhuang, M., Lei, F., Yang, B., & Zhao, M. (2013). GC/MS analysis of volatiles obtained by (MAXIM.) HARA leaf and stem. Food Chemistry, 136(2), 555-562.

IP

headspace solid-phase microextraction and simultaneous–distillation extraction from Rabdosia serra Li, S., Li, X., Wang, G., Nie, L., Yang, Y., Wu, H., Wei, F., Zhang, J., Tian, J., & Lin, R. (2012).

SC R

Rapid discrimination of Chinese red ginseng and Korean ginseng using an electronic nose coupled with chemometrics. Journal of Pharmaceutical and Biomedical Analysis, 70, 605-608. Mahattanatawee, K., & Rouseff, R. L. (2014). Comparison of aroma active and sulfur volatiles in three fragrant rice cultivars using GC-Olfactometry and GC-PFPD. Food Chemistry, 154, 1-6. Muskmelon

Juice

Characterized

by

Odor

NU

Pang, X., Chen, D., Hu, X., Zhang, Y., & Wu, J. (2012). Verification of Aroma Profiles of Jiashi Activity

Value

and

Gas

Chromatography –

Olfactometry/Detection Frequency Analysis: Aroma Reconstitution Experiments and Omission Tests.

MA

Journal Of Agricultural And Food Chemistry, 60(42), 10426-10432. Pang, X., Guo, X., Qin, Z., Yao, Y., Hu, X., & Wu, J. (2012). Identification of Aroma-Active Compounds in Jiashi Muskmelon Juice by GC-O-MS and OAV Calculation. Journal Of Agricultural And Food Chemistry, 60(17), 4179-4185.

D

Pino, J. A. (2014). Odour-active compounds in papaya fruit cv. Red Maradol. Food Chemistry, 146,

TE

120-126.

Pino, J. A., & Bent, L. (2013). Odour-active compounds in guava (Psidium guajava L. cv. Red Suprema). Journal Of The Science Of Food And Agriculture, 93(12), 3114-3120.

CE P

Pino, J. A., & Febles, Y. (2013). Odour-active compounds in banana fruit cv. Giant Cavendish. Food Chemistry, 141(2), 795-801.

Plotto, A., Margaria, C. A., Goodner, K. L., & Baldwin, E. A. (2008). Odour and flavour thresholds for key aroma components in an orange juice matrix: esters and miscellaneous compounds. Flavour And

AC

Fragrance Journal, 23(6), 398-406. Samykanno, K., Pang, E., & Marriott, P. J. (2013). Genotypic and environmental effects on flavor attributes of 'Albion' and 'Juliette' strawberry fruits. Scientia Horticulturae, 164, 633-642. Shao, Y., & He, Y. (2007). Nondestructive measurement of the internal quality of bayberry juice using Vis/NIR spectroscopy. Journal Of Food Engineering, 79(3), 1015-1019. Sun, S. Y., Jiang, W. G., & Zhao, Y. P. (2010). Characterization of the aroma-active compounds in five sweet cherry cultivars grown in Yantai (China). Flavour And Fragrance Journal, 25(4), 206-213. Wang, B., Xu, S. Y., & Sun, D. W. (2010). Application of the electronic nose to the identification of different milk flavorings. Food Research International, 43(1), 255-262. Wen, Y. Q., He, F., Zhu, B. Q., Lan, Y. B., Pan, Q. H., Li, C. Y., Reeves, M. J., & Wang, J. (2014). Free and glycosidically bound aroma compounds in cherry (Prunus avium L.). Food Chemistry, 152, 29-36. Xiao, Z., Yu, D., & Niu, Y. (2014). Characterization of aroma compounds of Chinese famous liquors bygas chromatography–mass spectrometry and flash GC electronic-nose. Journal of Chromatography B, 945-946, 92-100. Xu, Y. X., Zhang, M., Fang, Z. X., Sun, J. C., & Wang, Y. Q. (2014). How to improve bayberry (Myrica rubra Sieb. et Zucc.) juice flavour quality: Effect of juice processing and storage on volatile

ACCEPTED MANUSCRIPT compounds. Food Chemistry, 151, 40-46. Xuezhen, H., Jun, W., & Shanshan, Q. (2014). Authenticating cherry tomato juices-Discussion of different data standardization and fusion approaches based on electronic nose and tongue. Food Research International, 60, 173-179.

T

Zhou Shao-huan, Fang Zhong-xiang, Lü Yuan, Chen Jian-chu, Liu Dong-hong, & Ye Xing-qian (2009). Phenolics and antioxidant properties of bayberry (Myrica rubra Sieb. et Zucc.) pomance. Food

NU

SC R

IP

Chemistry, 112, 394-399.

MA

Figure captions

Fig.1. Effects of coating fibers (a), incubation time (b), extraction

D

temperature (c), and extraction time (d) on the peak areas for aroma

TE

volatiles in bayberry fruit by SPME-GC/MS.

CE P

Fig.2. PC1 vs. PC2 scatter plot of the main sources of variability among different bayberry cultivars (a) distinction between the samples (scores),

AC

BQ (a and b), DK (c and d), FHZ (f); (b) relationships between the volatile compounds (loadings). Codes correspond to the compounds described in Table 1. Fig.3. Two-dimensional (2D) PCA plots of different cultivars of bayberry fruits by electronic nose

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

Fig. 1

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

Fig. 2

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

Fig. 3

ACCEPTED MANUSCRIPT

RIa

IDb

Compounds

Aldehydes 816

MS, RI, AD

Hexanal

A2

859

MS, RI, AD

(E)-2-Hexenal

A3

901

MS, RI, AD

Heptanal

A4

915

MS, RI, AD

(E,E)-2,4-Hexadienal

A5

960

MS, RI

(E)-2-Heptenal

A6

1011

MS, RI, AD

Octanal

A7

1061

MS, RI, AD

(E)-2-Octenal

A8

1113

MS, RI, AD

Nonanal

A9

1169

MS, RI, AD

(E)-2-Nonenal

A10

1213

MS, RI, AD

Decanal

MA N TE D CE P

Alcohols

BQXJd

DKNHd

DKXJd

FHZd

20.39±2.25

19.14±1.45

19.53±1.14

21.78±0.17

14.41±1.07

1.26±0.11

2.37±0.27

3.38±0.3

3.96±0.46

1.3±0.44

0.34±0.03

0.25±0.01

0.25±0.04

nd

0.07±0.01

nd

0.43±0.06

0.83±0.12

0.59±0.12

nd

0.34±0.07

0.25±0.01

0.37±0.04

nd

nd

1.5±0.35

1.88±0.04

2.58±1.1

nd

nd

0.6±0.13

nd

0.19±0.02

nd

nd

3.73±0.26

2.87±0.45

3.16±0.52

2.11±0.47

1.16±0.63

0.61±0.16

0.91±0.02

nd

0.44±0.06

nd

0.41±0.05

0.32±0.04

0.43±0.05

0.39±0.03

0.09±0.05

29.2±2.88

28.41±0.87

30.72±0.71

29.27±0.95

17.02±2.38

nd

0.95±0.25

1.27±0.13

1.24±0.1

0.71±0.07

0.58±0.09

1.88±0.21

2.56±0.3

1.39±0.13

2.08±0.26

US

A1

concentration（mean ± SD）/µg/g c

BQCXd

CR

Code

IP

T

Table 1 The composition and concentration of volatile compounds identified in 5 different types of bayberry samples using SPME-GC/MS

862

MS, RI

(E)-3-Hexen-1-ol

B2

877

MS, RI, AD

1-Hexanol

B3

1109

MS, RI, AD

3,7-dimethyl-1,6-Octadien-3-ol

2.38±0.36

1.43±0.12

6.92±0.34

9.39±0.85

1.31±0.41

B4

1163

MS, RI, AD

(Z)-3-Nonen-1-ol

0.57±0.04

0.65±0.03

0.47±0.03

nd

0.08±0.14

B5

1175

MS, RI, AD

endo-Borneol

nd

0.48±0.02

nd

nd

0.44±0.1

B6

1187

MS, RI, AD

Terpinen-4-ol

0.35±0.11

0.44±0.02

nd

nd

0.26±0.07

B7

1200

MS, RI, AD

α-Terpineol

nd

nd

0.34±0.07

nd

1.98±0.79

B8

1206

MS, RI

(-)-Myrtenol

nd

nd

nd

nd

0.35±0.12

B9

1638

10,10-Dimethyl-2,6-dimethylenebicyclo[7.2.0]

3.39±0.14

2.06±0.25

7.44±0.51

5.8±1.34

nd

MS, RI

AC

B1

ACCEPTED MANUSCRIPT

Esters ＜800

MS, RI, AD

Ethyl Acetate

C2

933

MS, RI, AD

Hexanoic acid, methyl ester

nd

C3

941

MS, RI,

(Z)-3-Hexenoic acid, methyl ester

C4

1025

MS, RI

(E)-3-Hexen-1-ol, acetate

C5

1103

MS, RI, AD

Benzoic acid, methyl ester

C6

1180

MS, RI

Benzoic acid, ethyl ester

C7

1232

MS, RI, AD

3-Nonenoic acid, methyl ester

C8

1385

MS, RI

2-Propenoic acid, 3-phenyl-, methyl ester

18.99±1.3

17.82±2.38

7.21±1.69

0.83±0.06

0.33±0.01

nd

0.35±0.01

2.01±0.11

1.99±0.11

0.71±0.04

nd

nd

0.64±0.06

2.19±0.59

0.59±0.03

nd

nd

nd

0.33±0.01

nd

nd

4.69±0.19

6.39±0.11

0.32±0.06

0.35±0.04

nd

1.03±0.05

0.9±0.01

nd

nd

nd

0.82±0.36

0.35±0.02

0.46±0.04

nd

nd

0.47±0.13

0.36±0.1

0.47±0.15

nd

nd

7±0.47

11.47±0.32

6.08±0.68

1.65±0.1

0.35±0.01

nd

nd

nd

nd

42.46±1.28

nd

nd

nd

nd

0.4±0.07

nd

nd

nd

nd

1.77±0.33

nd

0.74±0.03

1.11±0.03

0.94±0.04

1.08±0.41

nd

nd

0.85±0.01

1.27±0.21

nd

US MA N

TE D

Terpenes

7.89±0.66

nd

CR

C1

IP

7.28±0.44

T

undecan-5.β.-ol

940

MS, RI, AD

α-Pinene

D2

955

MS, RI

Camphene

D3

983

MS, RI, AD

β-Pinene

D4

1036

MS, RI, AD

Limonene

D5

1047

MS, RI

trans-β-Ocimene

D6

1057

MS, RI, AD

(Z)- 3,7-dimethyl-1,3,6-Octatriene

0.63±0.37

nd

3.42±0.19

4.07±0.77

nd

D7

1067

MS, RI

γ-Terpinene

nd

0.52±0.05

nd

nd

nd

D8

1387

MS, RI

β-Bourbonene

0.35±0.05

nd

0.37±0.01

0.48±0.07

0.11±0.01

D9

1394

MS, RI

β-elemene

0.8±0.14

0.46±0.18

0.93±0.2

1.23±0.24

nd

D10

1409

MS, RI, AD

Isocaryophillene

1.05±0.12

0.89±0.18

3.53±0.92

8.14±1.27

nd

D11

1430

MS, RI, AD

Caryophyllene

108.97±3.05

73.99±2.58

168.83±5.95

234.8±2.52

3.07±0.16

D12

1430

MS, RI

Aromandendrene

0.5±0.06

nd

0.68±0.15

1.05±0.12

nd

D13

1452

MS, RI

Alloaromadendrene

0.56±0.09

nd

0.7±0.21

1±0.15

nd

AC

CE P

D1

ACCEPTED MANUSCRIPT

MS, RI

Humulene

4.25±0.52

2.41±0.81

6.02±1.2

7.54±1.44

nd

D15

1488

MS, RI

β-copaene

0.59±0.08

nd

nd

nd

0.41±0.04

D16

1492

MS, RI

β-selinene

0.75±0.11

nd

1.63±0.55

2.42±0.47

nd

D17

1509

MS, RI

α-Farnesene

0.59±0.14

nd

1.69±0.43

1.39±0.31

0.19±0.22

119.03±4.24

79.02±3.45

189.75±9.32

264.34±6.78

49.49±2.1

0.2±0.02

1.2±0.12

0.22±0

nd

nd

0.44±0.05

0.85±0.23

1.4±0.14

1.04±0.19

0.46±0.13

0.31±0.02

nd

nd

nd

nd

0.33±0.04

nd

0.77±0.04

0.37±0.04

0.07±0.02

nd

nd

1.75±0.12

1±0.19

nd

T

1466

CR

IP

D14

US

Others 892

MS, RI

Styrene

E2

968

MS, RI, AD

(E)-4-Oxohex-2-enal

E3

988

MS, RI, AD

1-Decen-3-one

E4

995

MS, RI, AD

6-methyl-5-Hepten-2-one

E5

1088

MS, RI

1-methyl-Pyrrolidine

E6

1139

MS, RI, AD

(E,Z)- 2,6-dimethyl-2,4,6-Octatriene

nd

nd

nd

0.43±0.12

nd

E7

1172

MS, RI, AD

Pinocarvone

nd

nd

nd

nd

0.43±0.15

E8

1374

MS, RI, AD

Dihydro-5-pentyl-2(3H)-Furanone

nd

0.4±0.02

nd

nd

nd

E9

1579

MS, RI

2,6-diethyl-Pyridine

1.49±0.19

nd

4.2±0.35

1.93±1.78

nd

E10

1601

MS, RI

Caryophyllene oxide

7.9±0.33

5.46±0.32

21.21±1.39

14.48±2.32

0.21±0.21

E11

1664

MS, RI

Isoaromadendrene epoxide

1.82±0.08

nd

4.51±0.31

3.46±0.63

nd

12.49±0.57

7.91±0.58

34.07±2.05

22.72±4.95

1.16±0.6

AC

CE P

TE D

MA N

E1

a

RI: retention indices

b

Volatiles were identified according to : MS, mass spectrum comparison using Wiley and NIST libraries; RI: retention index in agreement with literature value; AD, aroma description(odor).

c

µg/g: concentration was expressed in microgram per gram of juice, cyclohexanone as internal standard, and data listed were the mean of three assays ±SD( standard deviation)

d: BQCX: Biqi cultivar, CiXi, China: BQXJ: Biqi cultivar, XianJu, China; DKNH: Dongkui cutivar, NingHai, China; DKXJ: Dongkui cutivar, XianJu, China; FHZ: Fenhongzhong, ShangYu, China.

ACCEPTED MANUSCRIPT Table 2 Odor-active compounds identified in the headspace of bayberry juices by GC-O using detection frequency analysis (DFA)

RI

A1

816

A2

BQXJ

DKXJ

DKNH

FHZ

cut grass-like, green

8

8

8

8

8

859

green, apple

2

4

2

5

2

A3

901

fat, rancid

2

2

3

5

A4

915

green

A6

1011

lemon, green,fat, soap

A7

1061

green, nut, fat

A8

1113

fruit, flower, citrus, green

A9

1169

cucumber, green

A10

1213

oil, soap, orange peel

B2

877

resin, flower, green

B3

1109

flower,lavender

B4

1163

cucumber

B5

1175

camphor

B6

1187

turpentine, nutmeg

B7

1200

oil, mint

C1

＜800

pineapple

C2

933

fruit, fresh, sweet

C5

1103

herb, prune, lettuce

2

6

C7

1232

oil, fat

8

8

D1

940

pine, turpentine

6

D3

983

pine, resin, turpentine

8

D4

1036

lemon, citrus, mint

2

D6

1057

herb, ,mushroom

6

1409

wood

4

2

1430

wood,

spice

6

2

968

flower, sweet

8

8

E3

988

mushroom

4

E4

995

pepper, mushroom

6

E6

1139

grass, green

E7

1172

cucumber

E8

1374

sweet, spice

unknown

805

paint

unknown

972

acid

unknown

1020

turpentine, mint, spice

unknown

1029

solvent, gasoline, citrus

4

unknown

1389

lily, fat, citrus

2

E2

SC R 2

4

4

7 2

8

8

4

2

3

NU

MA

D

TE

2

8 2 6

CE P

AC

D11

IP

BQCX

D10

a

DFc

Odor descriptionb

T

Codea

4

4

8

8

2

5

5

2

6

2

5

5

4

4

6

2 2

3 2

4

3 2

2 2

6

7

2

3

4

6

2

4

4

5 2

2 4 2

4 4

2

2

6

2

3 2

Codes correspond to those in Table 1, unknown components were only identified by RI and odor (Others were identified by MS also)

b

Odor description at GC-sniffing port, odor description from http://www.flavournet.org.

c

Sum of times detected by eight assessors.

ACCEPTED MANUSCRIPT Highlights

1. GC-MS-O combined with E-Nose were used to analyze the

IP

T

aroma-active profile of Chinese bayberry.

SC R

2. PCA distinguished cultivars of bayberry samples successfully. 3. The performance of discrimination was confirmed by the combination of GC-MS-O and E-Nose.

AC

CE P

TE

D

MA

highest frequency by GC-O.

NU

4. β-caryophyllene is the predominant volatile and hexanal is with the