Chinese Traditional Pear Paste: Physicochemical Properties, Antioxidant Activities and Quality Evaluation

As a traditional folk medicine, pear paste has important nutritional and health effects. The physicochemical properties and antioxidant activities of pear pastes prepared from 23 different cultivars were investigated, including color parameters ( L*, a*, b* and h°), transmittance, pH, titratable acidity (TA), soluble sugar content, total phenolics content (TPC), total flavonoids content (TFC), DPPH and •OH radical scavenging activity (RSA), and ferric reducing antioxidant power (FRAP). It was demonstrated that the physicochemical properties and antioxidant activities of pear pastes from various cultivars differed significantly. Pear cultivars of “Mantianhong”, “Xiangshui” and “Anli” possessing higher TPC and TFC exhibited excellent antioxidant activity determined by DPPH RSA, •OH RSA and FRAP, while the lowest TPC and TFC was observed for the cultivars of "Xueqing", "Nansui", "Hongxiangsu", and “Xinli No. 7”, which also demonstrated the poor antioxidant activity. Multivariate analyses, including factor and cluster analysis, were used for the quality evaluation and separation of pear pastes based on their physicochemical and antioxidant properties. Factor analysis reduced the above thirteen parameters to final four effective ones, i.e. DPPH RSA, color b*, FRAP and TA, and subsequently these four parameters were used to construct the comprehensive evaluation prediction model for evaluating the quality of pear pastes. The pear pastes could be separated into three clusters and differentiated for the diverse of pear cultivars via cluster analysis. Consistently, “Mantianhong”, “Xiangshui” and “Anli” pear with higher quality clustered into one group, in contrast, "Xueqing", "Nansui", "Hongxiangsu", and “Xinli No. 7” with lower quality clustered into the other group. It provided a theoretical method to evaluate the quality of pear paste and may help the fruit processing industry select the more suitable pear cultivars for pear paste making.


Introduction
Pear is one of the main fruits in China with its cultivation area and production ranking the first in the world. The pear fruits have been proved to be rich in different kinds of bioactive polyphenol compounds, such as hydroxybenzoic acids, hydroxycinnamic acids, flavonoids and anthocyanins, varying with the cultivars [1][2][3]. Polyphenols have been proved to provide high nutritional and healthy benefits for humans with their antioxidant, anti-glycation and anti-inflammatory biological activities [4][5][6], and also play an important role in anti-aging, anti-cancer, and prevention of cardiovascular and neurological diseases [7][8][9].

Fruit Materials and Sample Preparation
Twenty-three cultivars of pear fruits bought from the local fruit wholesale market in Shijiazhuang were included in the study, see Table 1. The fruit samples were carefully selected for fruit size, absence of external and internal damage and without pests and diseases, washed and sliced into small pieces with the core and stem removed, and then crushed into juices with a juice presser (HU24FR3L, Hurom, Korea). After squeezing the juice and passing through 100 mesh sieve, the filtered pear juice was boiled with electric pottery stove (LC−EA3S, Guangdong Shunde Zhongchen Electric Co., Foshan, China) to make pear paste. The starting power of electric pottery stove was first set at 500 W, followed by gradient change of electric pottery stove power to 300 W and 200 W when the soluble solid content (SSC) of the concentrated pear juice reached 30 and 50 • Brix, respectively, until the SSC reached 70 • Brix, ending the boiling, and finally the concentrated pear juice become highly sticky pear paste. In order to detemine the physicochemical properties and antioxidant activities of pear paste, the original pastes have to be diluted until the SSC being 10 • Brix, since the pear paste was too sticky, and the obtained dilutions were used next. The firmness of the pear was monitored by hardometer (Top Instrument Co., Hangzhou, China). SSC ( • Brix) of pear juice was measured by a portable • Brix meter (Atago Co., Ltd., Tokyo, Japan). The pH of pear juice was monitored by a pH meter (ST3100, OHAUS (Changzhou) Instruments Co., China). TA of pear juice was determined by acid-base titration with slight modification according to the previous method [36], aliquots of 5 mL of pear juice placed into a 100 mL conical flask and titrated with standardized 0.01 M NaOH until phenolphthalein end point (pH 8.2 ± 0.1), and then the volume of NaOH was converted to g per 100 g of malic acid. The juice yield was calculated by dividing the weight of juice collected (kg) by the weight of pear fruit sample (kg), and the pear paste yield was calculated by dividing the weight of pear paste collected (kg) by the weight of pear fruit sample (kg).

Measurement of Color and Transmittance for the Diluted Pear Paste
A colour guide system (CR−400, Konica Minolta Co., Ltd., Osaka, Japan) with Illuminant D 65 and 10 • observer angle was used to measure the color (L*, a* and b*) of the diluted pear paste using a colorless and transparent petri dish filled with 20 mL diluted pear paste. L* indicated black (L* = 0) to white (L* = 100) component, a* indicated green (−) to red (+) component, and b* indicated blue (−) to yellow (+) component. The Hue angle value showed the variation of color which could be calculated by the formula: h • = arctan (b*/a*) with the variation between 0 • (purple−red) and 180 • (green), and the medium h = 90 • is for yellow. The transmittance (T) of the diluted pear paste was determined at 625 nm using a 1240 UV−vis spectrophotometer (Shimadzu Co., Ltd., Kyoto, Japan).
2.5. Measurement of TA, pH and Soluble Sugar Content for the Diluted Pear Paste TA and pH of the diluted pear paste were determined the same as the above pear juice. Soluble sugar content was analyzed by anthrone colorimetry according to a previous study [37], and glucose with concentration ranging between 0.02 mg/mL to 0.1 mg/mL was used as a standard curve.

Determination of Total Phenolics and Flavonoids Content for the Diluted Pear Paste
The total phenolics content (TPC) was determined according to the previous method reported by Jiang [13]. Briefly, 0.1 mL of the diluted sample in distilled water was mixed with 0.5 mL of 0.5 M Folin−Ciocalteu's reagent, 4.4 mL of distilled water and 1.0 mL of 7% sodium carbonate solution. The mixture was kept in dark for 2 hours at constant temperature of 30 • C in order to allow the reaction finish, and the absorbance was measured at 760 nm with a 1240 UV-vis spectrophotometer (Shimadzu Co., Ltd., Kyoto, Japan). The concentration of gallic acid ranging between 0.01 mg/mL to 0.05 mg/mL was used as a standard curve. Meanwhile, 0.1 mL of distilled water was used as blank control, and the TPC was calculated using the standard curve and expressed as milligram of gallic acid equivalent per 100 mL of the diluted pear paste.
The total flavonoids content (TFC) was measured according to the method described by Jiang [13]. Briefly, 0.2 mL of the diluted pear paste sample in distilled water was mixed with 0.3 mL of 5% NaNO 2 and 0.8 mL distilled water, and reacted at room temperature for 5 min. Afterwards, 0.3 mL of 10% AlCl 3 ·6H 2 O was added and the solutions were continued incubated for 5 min at room temperature before the addition of 2 mL of 1 M NaOH. The absorbance at 510 nm was measured after 15 min of incubation with a 1240 UV−vis spectrophotometer (Shimadzu Co., Ltd., Kyoto, Japan). The concentration of rutin ranging between 0.02 mg/mL to 0.1 mg/mL was used as a standard curve. Meanwhile, 0.1 mL of distilled water was used as blank control, and the TFC was calculated using a standard curve and expressed as milligram of rutin equivalent per 100 mL of the diluted pear paste.

Determination of DPPH Radical Scavenging Activity (DPPH RSA) for the Diluted Pear Paste
The DPPH RSA of the diluted pear paste sample was determined using the method described by Jiang with slight modification [13]. Briefly, 0.05 mL of diluted sample was mixed with 3.9 mL of 60 µM DPPH. After incubating the solution at room temperature in the dark for 30 min, the absorbance of the solution was measured at 517 nm. Meanwhile, 0.05 mL of 80% ethanol with 3.9 mL of 60 µM DPPH was used as blank control, The percentage of DPPH RSA was calculated according to the equation below: 2.8. Determination of • OH Radical Scavenging Activity ( • OH RSA) for the Diluted Pear Paste The • OH RSA of the diluted pear paste was determined using the previous reported method [38,39]. Briefly, 0.2 mL of diluted sample was mixed with 1 mL of 9 mM salicylic acid ethanol solution, 1 mL of 9 mM FeSO 4 solution and 0.8 mL of distilled water, and finally 1 mL of 8.8 mM H 2 O 2 solution was added to initiate the reaction. Simultaneously, the distilled water was used as a blank control. After the reaction at 37 • C for 30 min, both the sample solution and the blank control were centrifuged at 11,000 g for 6 min, and the absorbance of the supernatants of the diluted sample (A X ) and the blank control (A 0 ) was measured at 510 nm. Considering the different absorbance values of the diluted pear paste solution at 510 nm, 1 mL of 9 mmol/L FeSO 4 solution, 1 mL of 9 mmol/L salicylic acid ethanol solution, 0.2 mL of diluted sample and 1.8 mL of distilled water was measured as the background absorbance values (A X0 ). The percentage of • OH RSA was calculated according to the equation below: 2.9. Determination of Ferric Reducing Antioxidant Power (FRAP) Assay for the Diluted Pear Paste The FRAP of diluted pear paste was performed according to the method of Jiang [13] with slight modification. Briefly, 0.5 mL of diluted sample was mixed with 2 mL of PBS (pH~6.6) and 2 mL of 1% potassium ferricyanide and kept in water bath at 50 • C for 20 min. Then, 2 mL of 10% trichloroacetic acid aqueous solution was added and centrifuged for 10 min at 3000× g. Subsequently, 0.5mL supernatant was taken and mixed well with 0.4 mL of 0.1% ferric trichloride aqueous solution and 2 mL of distilled water and kept in dark for 30 min. The absorbance of the solution was measured at 700 nm and the reducing power of pear paste was expressed as absorbance units OD 700 .

Statistical Analysis
All samples were analyzed in triplicate and the results were expressed as the mean ± standard deviation. Spearman correlation analysis between the mean physicochemical properties and antioxidant activities of pear pastes prepared from 23 cultivars, Duncan's test for significance of difference, factor analysis and cluster analysis were all performed by Statistical Product and Service Solutions software SPSS 18 (IBM SPSS Statistics, Inc., Chicago, IL, United States) with the significant level set at p < 0.05.

Physicochemical Properties of Different Pear Cultivars
The single fruit weight and firmness of pear, TA, SSC and pH of pear juice, and juice and paste yields of 23 pear cultivars were seen in Table 1. The juice yields ranged between 69.57% to 88.09% with the two highest pear cultivars for "Xinli No. 7" (88.09%) and "Mantianhong" (86.45%), and the pear paste yields ranged between 8.17% to 13.16% with the two highest pear cultivars for "Shuihongxiao" (13.16%) and "xueqing" (12.73%).

Color Parameters and Transmittance of Diluted Pear Paste
Color is an important quality property of pear paste because it can determine its acceptability by the consumers. The color of pear paste prepared in the current study differed slightly depending on the cultivars which could be fully reflected by color parameters (L*, a*, b* and h

Color Parameters and Transmittance of Diluted Pear Paste
Color is an important quality property of pear paste because it can determine its acceptability by the consumers. The color of pear paste prepared in the current study differed slightly depending on the cultivars which could be fully reflected by color parameters (L*, a*, b* and h⸰) shown in Table 2. Among them, color a* is corresponded to the redgreen contribution, and color b* is corresponded to the yellow-blue contribution. All samples had a positive or slightly negative a* value (−1.16 to 4.91) which demonstrated that the samples were in red place, while all samples had positive b* values (5.53 to 14.64) meaning that all the samples were yellow. Furthermore, the h⸰ value of all samples ranged between 58.91 to 97.19, indicating the pear paste showed the overlapping color of red and yellow. Briefly, all the color parameters suggested the color of pear paste were of different level of reddish brown color. The transmittance of pear pastes at 625 nm ranging from 31.10% to 90.95% could reflect the clarification degree of the sample.  Table 2. Among them, color a* is corresponded to the red-green contribution, and color b* is corresponded to the yellow-blue contribution. All samples had a positive or slightly negative a* value (−1. 16

Color Parameters and Transmittance of Diluted Pear Paste
Color is an important quality property of pear paste because it can d ceptability by the consumers. The color of pear paste prepared in the cu fered slightly depending on the cultivars which could be fully reflected b ters (L*, a*, b* and h⸰) shown in Table 2. Among them, color a* is correspo green contribution, and color b* is corresponded to the yellow-blue contri ples had a positive or slightly negative a* value (−1.16 to 4.91) which de the samples were in red place, while all samples had positive b* value meaning that all the samples were yellow. Furthermore, the h⸰ value of all between 58.91 to 97.19, indicating the pear paste showed the overlapping yellow. Briefly, all the color parameters suggested the color of pear paste w level of reddish brown color. The transmittance of pear pastes at 625 nm 31.10% to 90.95% could reflect the clarification degree of the sample.

pH, TA and Soluble Sugar of Diluted Pear paste
The pH, TA and soluble sugar were analyzed from the processed pear pastes prepared from 23 different cultivars (Table 3). pH could be used to indicate the sourness of fruit juices, wines and other processed products in the quality evaluation [1,40]. The lowest values were found in "Anli" pear (pH 3.27) and "Xiangshui" pear (pH 3.40), while the two highest values were found in "Nansui" pear (pH 5.36) and "Hongxiangsu" pear (pH 5.31).
Pear paste made from different cultivars had lower pH, indicating that pear cultivar was sourer than the other studied cultivars. "Anli" pear rich in acid with the lowest pH was indeed sour, which was also called sour pear. Consistently, the highest TA was also found in "Anli" pear (0.652%), and in contrast, "Hongxiangsu" pear (0.068%) and "Nansui" pear (0.088%) had the first two lower TA, respectively. The soluble sugar content of the studied pear paste ranged between 59.47 to 84.63 mg/g.

Color Parameters and Transmittance of Diluted Pear Paste
Color is an important quality property of pear paste ceptability by the consumers. The color of pear paste pre fered slightly depending on the cultivars which could be ters (L*, a*, b* and h⸰) shown in Table 2. Among them, colo green contribution, and color b* is corresponded to the yel ples had a positive or slightly negative a* value (−1.16 to the samples were in red place, while all samples had p meaning that all the samples were yellow. Furthermore, th between 58.91 to 97.19, indicating the pear paste showed t yellow. Briefly, all the color parameters suggested the colo level of reddish brown color. The transmittance of pear 31.10% to 90.95% could reflect the clarification degree of t

pH, TA and Soluble Sugar of Diluted Pear paste
The pH, TA and soluble sugar were analyzed from pared from 23 different cultivars (Table 3). pH could be u fruit juices, wines and other processed products in the qu est values were found in "Anli" pear (pH 3.27) and "Xian two highest values were found in "Nansui" pear (pH 5.36 5.31). Pear paste made from different cultivars had lower p was sourer than the other studied cultivars. "Anli" pear

Total Phenolics and Flavonoids Content of Diluted Pear Paste
The results of total phenolics and flavonoids content (TPC and TFC) were shown in Table 3, which exhibited a very obvious difference related to pear cultivars. The TPC of pear pastes ranged between 10.52 to 51.98 mg/100 mL, with the two highest pear cultivars to be "Mantianhong" pear (51.98 mg/100 mL) and "Xiangshui" pear (43.22 mg/100 mL) and the two lowest to be "Nansui" pear (10.52 mg/100 mL) and "Xueqing" pear (11.30 mg/100 mL). Like TPC, the highest level of TFC was also found in "Mantianhong" pear (42.28 mg/100 mL) and "Xiangshui" pear (20.44 mg/100 mL), so was the lowest level as "Nansui" pear (2.74 mg/100 mL) and "Xueqing" pear (3.13 mg/100 mL). According to the previous reports, phenolic profiles of pear were complex and the content of individual phenolic compounds also differed significantly in different pear cultivars [1][2][3]. The total polyphenolic content ranged differently for five typical pyrus species including ten pear cultivars [2]. Additionally, the contents and types of phenolic compounds in different European pear cultivars also differed significantly [1,3]. It may due to the discrepancies of the types and contents of phenolic compounds in different pear cultivars that led to the differences in TPC and TFC of pear paste.
It was confirmed that the phenolics and flavonoids played a very important role in antioxidant activities, which might ascribed to hydrogen atom transfer ability of their phenolic hydroxyl groups and the stability of the formed phenoxy radicals [41]. As different pear cultivars contained different phenolics and flavonoids profile, thus the antioxidant activities of pear pastes differed significantly. Indeed, the antioxidant activities of pear pastes were found to be significantly correlated with the TPC and TFC in pear paste in our study, which was also reported in other studies [11,42]. Figure 1 showed the heat map of Spearman correlation coefficient for the physicochemical properties and antioxidant activities of 23 pear pastes. With the correlation significant level setting at p < 0.05, highly significant positive correlations were observed between the TPC (r = 0.88) and TFC (r = 0.86) with DPPH RSA, and highly significant positive correlations were also observed between the TPC (r = 0.76) and TFC (r = 0.71) with • OH RSA. The moderate significant correlations were observed between the TPC (r = 0.47) with FRAP, however exception was observed for no significant correlations between the TPC (r = 0.39) with FRAP. Furthermore, we also found a highly significant correlation between the pH and TA of pear pastes with their antioxidant activities determined by DPPH RSA and • OH RSA, shown in Figure 1, which suggested that organic acids were also closely related to the antioxidant activities of pear pastes. It has already been confirmed that not only polyphenols but also organic acids have good antioxidant activities and could be the good dietary source of antioxidants. For example, phenolic compounds and organic acids might be responsible for the antioxidant activities of fruit vinegars and Camellia oleifera cake according to the previous reports [43,44]. In addition, organic acids showed a synergistic effect with α-terpinene in DPPH scavenging activity [45].

Evaluation Method of the Quality of Diluted Pear Paste
The physicochemical and antioxidant properties of each pear paste sample were performed by multivariate analyses, including factor and cluster analysis, in order to construct the comprehensive evaluation prediction model for the quality of pear pastes and to separate the pear pastes into different groups.
Before the application of factor analysis, it was necessary to carry out the Kaiser−Meyer-Olkin (KMO) and Bartlett ball type test. KMO test was used to test the correlation between variables, if the KMO test was below 0.5, the application of factor analysis was not suitable on this occasion. While Bartlett ball type test was used to test whether the correlation matrix was a unit matrix, and the use of factor analysis should be performed carefully if the test p > 0.05 [27]. Based on our current research, KMO test was 0.679 greater than 0.5 and p < 0.001 in Bartlett's test was less than 0.05, both indicating the data was suitable for factor analysis. Factor analysis was firstly used to seek the principal factor affecting the quality of pear pastes from 23 cultivars. A Varimax-rotation was performed in order to ensure that the abstracted principal factors were uncorrelated and facilitated the interpretation of the results. By taking eigenvalues of greater than 0.7, four principal factors (F1, F2, F3 and F4) were extracted. Table 5 showed the loadings, eigenvalues, percent of variance, cumulative variance and principal factor weight obtained from the factor analysis. The four principal factors explaining 42.4%, 25.1%, 14.6% and 8.4% of the total variance, respectively, with the cumulative variance to be 90.5%, could contain all sufficient information. According to the loading values and only taking those absolute values p ≥ 0.7 into consideration, pH, TA, DPPH RSA and • OH RSA were the main variables correlating to F1. TFC and FRAP were correlated to F2. Color b* was correlated to F3 and soluble sugar was the main variable correlating to F4. Then the score of the comprehensive factor (F) for all the pear paste samples could be calculated according to the score of each principal factor of pear paste sample and the corresponding principal factor weight, as the equation be: (2) our study, which was also reported in other studies [11,42]. Figure 1 showed the heat map of Spearman correlation coefficient for the physicochemical properties and antioxidant activities of 23 pear pastes. With the correlation significant level setting at p < 0.05, highly significant positive correlations were observed between the TPC (r = 0.88) and TFC (r = 0.86) with DPPH RSA, and highly significant positive correlations were also observed between the TPC (r = 0.76) and TFC (r = 0.71) with • OH RSA. The moderate significant correlations were observed between the TPC (r = 0.47) with FRAP, however exception was observed for no significant correlations between the TPC (r = 0.39) with FRAP. Furthermore, we also found a highly significant correlation between the pH and TA of pear pastes with their antioxidant activities determined by DPPH RSA and • OH RSA, shown in Figure  1, which suggested that organic acids were also closely related to the antioxidant activities of pear pastes. It has already been confirmed that not only polyphenols but also organic acids have good antioxidant activities and could be the good dietary source of antioxidants. For example, phenolic compounds and organic acids might be responsible for the antioxidant activities of fruit vinegars and Camellia oleifera cake according to the previous reports [43,44]. In addition, organic acids showed a synergistic effect with α-terpinene in DPPH scavenging activity [45].

Evaluation Method of the Quality of Diluted Pear Paste
The physicochemical and antioxidant properties of each pear paste sample were performed by multivariate analyses, including factor and cluster analysis, in order to construct the comprehensive evaluation prediction model for the quality of pear pastes and to separate the pear pastes into different groups.  Based on this, pear pastes prepared from 23 cultivars could be ranked by the score of the comprehensive factor F ( Table 6), demonstrating that the higher the score of F, the better the quality of pear paste. Among them, the top three were "Mantianhong" pear, "Xiangshui" pear and "Anli" pear, while the last three were "Xinli No. 7" pear, "Hongxiangsu" pear and "Nansui" pear, in good agreement with the the top three and bottom three of the results of the antioxidant activities.

Color Parameters and Transmittance of Diluted Pear Paste
Color is an important quality property of pear paste because it can determine its acceptability by the consumers. The color of pear paste prepared in the current study differed slightly depending on the cultivars which could be fully reflected by color parameters (L*, a*, b* and h⸰) shown in Table 2. Among them, color a* is corresponded to the redgreen contribution, and color b* is corresponded to the yellow-blue contribution. All samples had a positive or slightly negative a* value (−1.16 to 4.91) which demonstrated that the samples were in red place, while all samples had positive b* values (5.53 to 14.64) meaning that all the samples were yellow. Furthermore, the h⸰ value of all samples ranged between 58.91 to 97.19, indicating the pear paste showed the overlapping color of red and yellow. Briefly, all the color parameters suggested the color of pear paste were of different level of reddish brown color. The transmittance of pear pastes at 625 nm ranging from 31.10% to 90.95% could reflect the clarification degree of the sample.

pH, TA and Soluble Sugar of Diluted Pear paste
The pH, TA and soluble sugar were analyzed from the processed pear pastes prepared from 23 different cultivars (Table 3). pH could be used to indicate the sourness of fruit juices, wines and other processed products in the quality evaluation [1,40]. The lowest values were found in "Anli" pear (pH 3.27) and "Xiangshui" pear (pH 3.40), while the two highest values were found in "Nansui" pear (pH 5.36) and "Hongxiangsu" pear (pH 5.31). Pear paste made from different cultivars had lower pH, indicating that pear cultivar was sourer than the other studied cultivars. "Anli" pear rich in acid with the lowest pH  The comprehensive evaluation prediction model for the quality of pear paste was afterwards constructed according to the multiple linear and stepwise regression, in which the above score of F was taken as the dependent variable Y, and the eight physicochemical parameters (pH, TA, DPPH RSA, • OH RSA, TFC, FRAP, b* and total soluble sugar) correlated to the four principal factors were the independent variable X. And finally, the prediction model containing only four effective parameters was obtained and expressed as: F = −2.084 + 0.015DPPH RSA + 0.114b* + 0.357FRAP + 1.408TA with r = 0.994, p < 0.001. In which, F was the comprehensive evaluation prediction factor. The Variance inflation Factor (VIF) for the four effective parameters DPPH RSA, b*, FRAP and TA were 3.600, 1.164, 1.259, and 3.492 respectively, indicating the less multi−collinearity and no overfitting in this linear equation. The partial correlation coefficient was 0.958, 0.974, 0.912, and 0.869 for the four effective parameters separately, suggesting that the larger the absolute value, the greater its role in the comprehensive evaluation prediction model for the quality of pear paste. According to the equation, the scores of F (Table 6) could be calculated if we substitute the values of four effective parameters into the Equation (2). With F and F were highly significant positive correlation by correlation analysis, we got the comprehensive evaluation prediction model with only four effective parameters, which could evaluate the quality of pear paste. Cluster analysis was then performed in order to test the similarity and search for groupings among the different pear paste samples based on the calculated score of F . Three clusters were found at a similarity level of 3 using Euclidean Distance (Figure 2). The first cluster included "Xiangshui" pear, "Mantianhong" pear, and "Anli" pear, with their F score ranging from 0.777 to 1.154, exactly to be the top three with the highest F score. The second was made up of the medium quality of pear paste, including pear cultivars for Red pear, "Qiubai" pear, "Redzaosu" pear, "Shuihongxiao" pear, "Zaomisuan" pear, "Nanguo" pear, "Akiziki" pear, "Yali" pear and "Huagai" pear, with their F scores from −0.034 to 0.539. The other eleven pear paste samples, with their F scores on the scale of −0.834 to −0.245, clustered together. It was necessary to note that the last group splited into two new clusters at a similarity level of 1, in which "Hongxiangsu" pear, "Xueqing"pear, "Nansui" pear, and "Xinli No. 7" pear were contained in one of the new clusters, just the same as in the bottom four of the lowest F score, which indicated that cluster analysis using F scores was accurate. into two new clusters at a similarity level of 1, in which "Hongxiangsu" pear, "Xueqing"pear, "Nansui" pear, and "Xinli No. 7" pear were contained in one of the new clusters, just the same as in the bottom four of the lowest F score, which indicated that cluster analysis using F′ scores was accurate.

Conclusions
It was concluded that the quality of pear paste was closely related to the pear cultivar through the measurement of the physicochemical properties and the antioxidant activities of different pear paste samples. The comprehensive evaluation model for the quality of pear paste was constructed based on only four effective parameters, i. e. b*, TA, DPPH RSA and FRAP. According to the score of F and F′, pear cultivars of "Mantianhong",

Conclusions
It was concluded that the quality of pear paste was closely related to the pear cultivar through the measurement of the physicochemical properties and the antioxidant activities of different pear paste samples. The comprehensive evaluation model for the quality of pear paste was constructed based on only four effective parameters, i.e. b*, TA, DPPH RSA and FRAP. According to the score of F and F , pear cultivars of "Mantianhong", "Xiangshui", and "Anli" with higher phenolics and flavonoids content and higher antioxidant activities ranked the top three, conversely, pear cultivars of "Xueqing", "Nansui", "Hongxiangsu", and "Xinli No. 7" with lower phenolics and flavonoids content and lower antioxidant activities ranked the bottom four, and both of them clustered into separate groups respectively, which demonstrating that the model could evaluate the quality of pear paste correctly, and cluster analysis could differentiate the pear paste samples successfully based on calculated score of F . Therefore, the research shed light on the physicochemical and antioxidant properties of pear paste from different cultivars and established an evaluation method for the quality of pear paste.