Analysis of Taste Compounds in Red Sufu during Fermentation and Discrimination with Electronic Tongue

WANG Peng, WANG Wenping,, XU Dandan, ZHANG Xin, GAO Hang, ZHANG Jian, SUN Yong, WANG Fenghuan

(1.Beijing Academy of Food Sciences, Beijing 100068, China;2.Beijing Laboratory of Food Quality and Safety, Beijing Technology and Business University, Beijing 100048, China)

Abstract: In order to evaluate the taste compounds and taste characteristics of red sufu during its fermentation process,the samples collected at different fermentation stages were analyzed by ion chromatography, amino acid analyzer, and electronic tongue.The obtained data were analyzed by principal component analysis (PCA) and hierarchical cluster analysis(HCA).Results showed that a total of 17 free amino acids (FAAs), 6 organic acids and 8 taste attributes were identified and quantified in all samples.Total contents of FAAs increased significantly at pre-fermentation and post-fermentation stages,while total contents of organic acids showed a significant increase at the pre-fermentation stage and the middle stage of post-fermentation (P < 0.05).The levels of most FAAs and organic acids increased significantly at the pre-fermentation and post-fermentation stages (P < 0.05).Electronic tongue analysis showed that umami, saltiness, bitterness and richness were major taste attributes.Furthermore, correlation analysis between taste compounds and taste attributes showed that free amino acids (Glu, Asp, Ala, Gly, His, Cys, Phe, Met and Val), organic acids (lactic acid, acetic acid, succinic acid) and some other chemical components (reducing sugar, amino acid nitrogen and NaCl) played an important role in the taste characteristics of red sufu.The results of PCA and HCA showed that the taste compounds and taste attributes could be used to evaluate its maturity extent and that electronic tongue could be a rapid tool for the taste evaluation of red sufu.

Keywords: red sufu; electronic tongue; taste compounds; free amino acids; organic acids

Sufu, existing in China with a long history for more than 1 000 years, is a soft, creamy and highly flavored fermented soybean curd product[1].Sufu has been widely consumed as an appetizer and side dish in China and other Asian countries,attributing to its unique flavor, delicious taste, and remarkable nutritional value[2].It has some functional components, such as soybean oligopeptide[3], isoflavone[4], and oligosaccharide[5].In terms of the color and flavor, sufu can be categorized into four types: red sufu, white sufu, grey sufu and others[6].Red sufu, which has a bright red color and unique flavor, gains more preference of consumers and takes a large proportion of gross sales.Currently, there are numerous red sufu products available in China.The qualities of red sufu are affected by several factors, such as fermentation conditions, microbes and raw materials used by manufacturers[7].Among them,Wangzhihe red sufu is the representative of northern region of China.

The production process of red sufu comprises four steps, including the formation of soybean curd (tofu),phetze prepared byActinomucor elegansfermentation (prefermentation), salting phetze, and ripening of sufu (postfermentation) (Fig.1)[8].During fermentation, its qualitative characteristics vary in terms of the metabolites present,due to enzymatic activities and physiochemical reactions occurred in different stages[9-10].Complex macromolecules are decomposed to smaller nutrients easily absorbed, such as free amino acids (FAAs), small peptides, reducing sugars and free fatty acids[11].These substances contribute nutritive value and taste characteristics to red sufu products.

Fig.1 Flow chart of red sufu fermentation with different sampling points

Taste is an important sensory property of red sufu, along with aroma and texture[2].In addition, it could determine consumers’ acceptance and selection directly[12].The unique taste of red sufu is mainly constituted by five basic tastes including umami, saltiness, bitterness, sweetness,and sourness, among which the umami and saltiness taste are particularly prominent[13].Different manufacturing processes, basic ingredients, and microbial communities will significantly affect their qualitative characteristics with remarkable differences in metabolites, which in turn determine the discrepancies of red sufu characteristic taste[9].Metabolites, especially FAAs and organic acids, have been found to be not only bioactive compounds but also tasteactive substances in fermented products[12,14-16].Previous studies confirmed that Pro and Pro-containing peptides are the main bitterness taste contributors[12], while hydrophobic(Val, Leu, Ile, Phe, Tyr, Trp, and Met) and basic (His and Arg) FAAs contribute to the bitterness[17].Acidic amino acids (Glu and Asp), Ala and Glu-enriched oligopeptides can provide umami taste for fermented products[18].Lactic acid and acetic acid are main contributors to sourness taste[15,19-20].Besides, the composition of organic acids with harmonious proportion affects the taste profile of red sufu, especially sour,bitterness as well as astringency[21].However, up to now,a few studies have been reported to investigate FAAs and organic acids of sufu products[2,6,11,22-23], but the relationship between metabolites and taste characteristics has not been discussed.The contribution to the taste characteristics cannot be evaluated in terms of a single compound[20].

As a technology to assist methods used for analyzing taste-active compounds, taste sensing system (electronic tongue) is regarded as an appropriate technique, since sensory evaluation is expensive, time consuming and confined to some conditions[24].As an alternative for human senses,the taste sensing system has been widely used for sensory analysis to measure the quality of foods, and obtain globalinformation about samples using a simple, non-specific sensor array and pattern-recognition software system[25].With the advantages of simpleness, rapidness, portability and high repeatability, it has gain remarkable performance for the taste evaluation[26], process monitoring[27], food authentication[28],and quantitative analysis[29]in many products.In recent years,few studies have been conducted to investigate the tastes of red sufu by taste sensing system.Xie Chunzhi et al.[2].analyzed taste characteristics and taste-active substances of three categories of sufu products, showing that different sufu categories were distinguished by taste sensing system.To the best of our knowledge, the real-time change of taste characteristics in red sufu during fermentation has not been investigated.In addition, research on the correlation between taste sensing system and liquid chromatographic analysis of red sufu during fermentation has not been found.Furthermore, the current Chinese sufu industrial standard(SB/T 10170-2007) mainly consists of sensory evaluation,physiochemical indicators, and hygienic requirements as the basis of quality evaluation.Assessment of red sufu taste indexes may be considered as a complementary technique for its quality evaluation.

The aim of study is to reveal the correlations between the chemical compositions and taste characteristics of red sufu during fermentation process.The techniques of liquid chromatography (LC), ion chromatography (IC), and taste sensing system were adopted to measure the dynamic change of amino acids, organic acids as well as taste attributes of sufu samples at different fermentation stages.This study evaluated taste changes of red sufu during fermentation process by chemical analysis of the taste compounds and taste sensing system.The results will be useful for the quality evaluation of red sufu and monitoring taste changes during fermentation process.

1 Materials and Methods

1.1 Materials and reagents

Samples at different fermentation stages were collected in three batches from Beijing Wangzhihe Food Group Co.Ltd., including pre-fermentation stage: days 0 (S1), 1 (S2)and 2 (S3); Salted phetze stage: days 2 (S4), 4 (S5) and 6(S6); Post-fermentation stage: days 0 (S7), 5 (S8), 15 (S9),30 (S10), 60 (S11) and 90 (S12).The process flow chart is illustrated in Fig.1.

Lactic acid (> 99%), citric acid (> 99%), sodium acetate(> 99%), sodium formate (> 99%), tartaric acid (> 98%),oxalic acid (> 99%), and succinic acid standards (> 99%)China National Standard Material Center (Beijing, China);Amino acids standard solution (2.5 mmol/L) Wako Pure Chemical Industries, Ltd.(Osaka, Japan); HCl (GR grade)Sigma-Aldrich Co.(St.Louis, USA); The derivatives and mobile phase solution Hitachi High-Technologies Co.(Tokyo, Japan).

1.2 Instruments and Equipments

AH-30 high-speed homogenizer RayKol Group Co.Ltd., China; 5424R refrigerated centrifuge Eppendorf Inc.,Germany; L-8900 automatic amino acid analyzer Hitachi High-Technologies Co.Japan; TS-5000Z taste sensing system Insent Inc., Japan; ICS 5000+Ion chromatography,equipped with conductivity detector Thermo Fisher Scientific Inc., USA.

1.3 Methods

1.3.1 Measurement of FAAs

0.20 g homogenous sufu sample was mixed with 20 mmol/L HCl solution in a 10 mL volumetric flask and sonicated for 30 min at room temperature.And then, they were centrifuged at 12 000 ×gfor 5 min.The supernatant was filtered through a 0.22 μm filter membrane, and then FAAs were determined by Hitachi L-8900 automatic amino acid analyzer (Tokyo, Japan).Chromatographic conditions were performed according to the method described by Wu Junrui et al.[30].

1.3.2 Measurement of organic acids

Analysis of organic acids were conducted based on Chinese industry standard SN/T 4675.5-2016 (Determination of organic acids in wine for export-ion chromatography)using a Dionex IC-5000+IC system (Thermo Scientific,Sunnyvale, CA) equipped with analytical gradient pumps,conductivity detectors (CD), an AERS_4 mm Dionex anion self-regenerating suppressor and an AS-AP auto-sampler.Briefly, weigh 0.20 g homogenous sufu samples in 10 mL volumetric flask and mixed with deionized water to volume.Sonicate for 30 min and centrifuge for 5 min at 12 000 ×g.Filter the supernatant into injection bottles through a 0.22 μm filter membrane.Dionex IonPac AS11-HC (4 mm × 250 mm,Thermo Scientific) with a guard column IonPac AG11-HC(4 mm × 50 mm, Thermo Scientific) was performed to separate organic acid anions and the eluent was potassium hydroxide (KOH) with a flow rate of 1.0 mL/min.The optimized KOH gradient elution program was as follows:0–5.0 min, 1 mmol/L; 40.0–45.0 min, 36 mmol/L;45.1–50.0 min, 1 mmol/L.The column temperature was 30 ℃ and the injection volume was 25 μL.

1.3.3 Measurement of taste characteristics by taste sensing system

The taste intensity of sufu samples was analyzed by taste sensing system TS-5000Z (Insent Ltd., Japan).It is composed of 2 reference electrodes, 5 sensor probes, an auto-sampler,an electronic unit for data acquisition and a computer with chemometric software[25].20.00 g Homogenous red sufu sample was transferred in a 100 mL volumetric flask and mixed with deionized water to volume.The solution was sonicated for 30 min and centrifuged at 12 000 ×gfor 5 min.The supernatant was filtered with the filter paper.In addition,a tasteless sample, comprised of 30 mmol/L KCl and 0.3 mmol/L tartaric acid, was used as the reference.Each sample was measured 3 times and the data was collected for further analysis.

1.4 Statistics analysis

All experiments were performed in triplicate and the data was standardized before analysis.Significant differences(P< 0.05) were performed by one-way ANOVA using SPSS 20.0 (SPSS, Chicago, IL, USA).Chemical compounds and taste attributes during red sufu fermentation were analyzed with principal component analysis (PCA) by R software using FactoMineR and Factoextra.The metabolites were visualized by hierarchical cluster analysis (HCA) with pheatmap package, as well as the correlation of taste attributes identified by correlation matrix with Hmisc package.In addition, the possible relationships between metabolites and taste intensities were explored by using R software with the mixOmics package.

2 Results and Analysis

2.1 Physiochemical properties and metabolites in red sufu during fermentation

Physiochemical properties of red sufu samples during fermentation, including amino acid nitrogen (AAN), NaCl,and reducing sugars has been discussed in our previous study[8].AAN had sharply increased at pre-fermentation stage, ranging from 0.01 g/100 g to 0.65 g/100 g, decreased to 0.26 g/100 g at S7 and gradually increased to 0.59 g/100 g at post-fermentation stage.NaCl significantly increased to 13.6 g/100 g at salted phetze stage and decreased after the addition of dressing mixture (8.02 g/100 g), it remained stable at post-fermentation stage.Reducing sugars content showed a rapid increase after the addition of dressing mixture and achieved the maximum value (10.20 g/100 g) at S10, then appearing a decreased tendency until the end of post-fermentation.

FAAs were reported to be contributors to the taste of fermented soybean food, also identified as major nitrogen compounds during fermentation[23].As demonstrated in Fig.2a, a total of 17 FAAs, including 7 essential amino acids, was identified in red sufu samples during fermentation.Glutamic acid (Glu), alanine (Ala), leucine (Leu), phenylalanine(Phe), and lysine (Lys) were identified as the major FAAs.Generally, the content of total FAAs and total essential amino acids (EAAs) increased significantly (P< 0.05) at pre-fermentation and post-fermentation stages (Fig.2b),which is consistent with those presented in fermented soy products[10-11,31].The increase of total FAAs is mainly due tothe action of proteases released by microorganisms in sufu[23].A significant decrease (P< 0.05) was only observed at the former stage of salted phetze, due to the interaction between brine solution and phetze.However, slight differences(P> 0.05) were observed at the latter stage of postfermentation stage (S11–S12) in total FAAs and total EAAs.

Fig.2 Variation in contents of FAAs and organic acids in red sufu samples during fermentation

Taste perception, one of the important quality indexes,is an important part in food industries[32].According to their taste attributes, FAAs could be classified into four categories,including 9 bitter FAAs (methionine (Met), isoleucine (Ile),Leu, Phe, Lys, histidine (His), arginine (Arg), tyrosine (Tyr),and valine (Val)), 5 sweet (glycine (Gly), Ala, threonine(Thr), serine (Ser), and proline (Pro)), 2 umami (aspartic acid (Asp) and Glu) and 1 tasteless (cysteine (Cys))[15].According to Table S1, bitterness and umami were the major groups, which is in accordance with other fermented soybean products[17].As the major bitter FAAs, Leu, Phe, and Lys are mainly generated from unbalanced proteolysis and peptide hydrolysis.They showed the same tendency, which increased at pre-fermentation and post-fermentation stages.Glu, the major umami FAA identified in fermented food,resulted from proteolysis or conversion of glutamine by glutaminase[12].It showed significant increases (P< 0.05) at pre-fermentation and post-fermentation stages.Ala, identified as the main sweet FAA, increased at pre-fermentation stage and remained relatively stable at post-fermentation stage.As the total FAAs content increased during post-fermentation,the relative proportion of each group remained relatively constant (Table 1).

Table 1 FAA profiles of red sufu samples during fermentation mg/kg

Organic acids are metabolites of lactic acid bacteria during fermentation process[11].The composition of organic acids produced during fermentation depends on the species of organisms, culture composition and growth conditions[16].According to Fig.2c, six organic acids were identified in sufu samples during fermentation, including formic, acetic,lactic, oxalic, citric and succinic acids.As demonstrated in Fig.2d, significant increases (P< 0.05) of total organic acids existed at pre-fermentation stage, S5–S7, as well as S8–S11,which might be caused by the increase ofLactococcus[11].In addition, significant decreases (P< 0.05) were investigated at S3–S5, S7–S8 and S11–S12, and this may attribute to the addition of brine solution and dressing mixture, as well as the esterification reaction[32].

As shown in Fig.2c, among six organic acids identified in red sufu, acetic acid and lactic acid were the dominant organic acids.They showed significant increases (P< 0.05)at pre-fermentation stage, which were possibly produced by lactic acid bacteria, such asLactococcusandWeissela, which had been identified at pre-fermentation stage[11].A rapid decrease was investigated in S8, which could be reasoned by the production of ethyl esters in the presence of yeast and abundant ethanol[32].After that, their levels increased inversely with the reduction of reducing sugars, suggesting that the accumulation of organic acids depends on sugar metabolism.Their content had a rapid reduction at the end of fermentation, which is also investigated in Cheonggukjang[31].Oxalic acid and succinic acid, as intermediates in tricarboxylic acid (TCA) cycle, decreased gradually at postfermentation stage[9].Beyond that, succinic acid showed significant increase at pre-fermentation stage, and oxalic acid increased significantly at salted phetze stage.Citric acid tended to decrease at pre-fermentation and post-fermentation stages, and it could be explained by the conversion to formic acid and acetic acid[31].Formic acid, possibly related to the conversion of metabolic pathway from homolactic to mix acids fermentation[33], fluctuated during fermentation and achieved the maximum amount at S6.

2.2 Taste attributes of red sufu during fermentation

Red sufu was subjected to taste sensing system during different fermentation stages, and intensities of eight taste attributes were shown in Fig.3a, including sourness,astringency, bitterness, saltiness, umami, richness, astringency aftertaste (aftertaste-A) and bitterness aftertaste (aftertaste-B).According to the tasteless blank sample, at which the tasteless points for taste attributes were 0, except for saltiness (-6) and sourness (-13), taste attributes with intensities above tasteless points were considered meaningful.Therefore, a new radar plot (Fig.3b) was established based on meaningful taste attributes, including umami, saltiness, bitterness, astringency,aftertaste-A, aftertaste-B and richness.As exhibited in Fig.3b, umami, saltiness, bitterness, and richness were dominant taste attributes with the response values significantly higher than those of other tastes during fermentation.Umami increased at pre-fermentation stage and initial stage of postfermentation (S8–S10), due to the increase of FAAs[6].Saltiness had increased 10 times at salted phetze stage, and this is positively correlated with the addition of NaCl in the brine solution.While, bitterness increased at pre-fermentation stage and the prophase of post-fermentation stages (S7–S10),then decreased dramatically along with post-fermentation progress.Even though the content of umami FAAs in S6 was less than that in S3 with bitterness FAAs similar in both samples, umami and bitterness attributes showed an increasing and decreasing trend, respectively.Changes of bitterness and umami intensities would be explained by the interaction between taste-active amino acids and salt, which NaCl could mask bitterness and enhance umami taste[12].The response value of richness, which was not meaningful at prefermentation and salted phetze stages, presented an increasing trend and remained stable at post-fermentation stage.Whereas, astringency, aftertaste-A and aftertaste-B, aversive to most consumers, decreased at the last two months of postfermentation stage.These results exhibited that fermentation could offer a better taste to soybean products with the reinforcement of pleasant tastes, as well as the reduction of undesirable attributes.

Fig.3 Radar plots for taste sensing system scores in terms of overall taste (a) and meaningful taste attributes (b) of red sufu samples during fermentation

2.3 Principle component analysis of physiochemical properties, metabolites and tastes attributes

PCA is a multivariate statistical analysis to simply analyze by exhibiting similarities and differences among samples by reducing the number of dimensions without much loss of information[34].In this study, PCA analysis was performed to evaluate the differences among physiochemical properties and chemical compositions for samples at different fermentation stages.As shown in the score plot (Fig.4a),the first principal component (PC1) and second principal component (PC2) accounted for 46.9% and 25.4% of the total variance, respectively.Two separated clusters, representing pre-fermentation (S1–S3) and salted phetze stages (S4–S6), as well as post-fermentation stage (S7–S12), were shown on the negative and positive axis of PC1, respectively.It indicated that PC1 can distinguish samples at different fermentation stages and remarkable differences of chemical compositions existed in these samples, probably due to the inoculation ofMocur, addition of brine solution, as well as long term fermentation in dressing mixture, respectively.Furthermore,samples at pre-fermentation stage had larger distances,indicating that the metabolism ofMocurhad important effects on the chemical compositions of red sufu.In addition, large distances were observed between samples at salted phetze stage, which demonstrated that addition of brine solution also played important role on the chemical compositions of red sufu.However, samples at post-fermentation stage were clustered in the same region and the two groups are nearly merged between S11 and S12, indicating that their chemical compositions became gradually stable as post-fermentation processes and the fermentation of red sufu could be finished in 3 months.

Fig.4 PCA plots for physiochemical properties, metabolites and taste attributes of red sufu

PCA loading plot was generated to identify these attributes responsible for the differentiation of samples in Fig.4b.The longer the arrows in the loading plot, the more important they are for the separations of samples during fermentation[35].The positive PC1 axis was greatly affected by most FAAs, which increased gradually during fermentation,such as Glu, Val, His, Ile, Met, and Leu.Therefore, these substances can be used to indicate the post-fermentation stage.On the other hand, the negative PC1 axis was influenced by citric acid and oxalic acid, while Pro, succinic acid, Ala, formic acid and NaCl were strongly correlated with the positive PC2.They will be considered as indicators of samples at prefermentation and salted phetze stages.

To validate the tastes changes during the whole fermentation process, PCA was generated in terms of taste intensities (Fig.4c).As evidenced by PCA, the score plot uncovered that samples were significantly different from each other on PC1, with respect to their taste intensities.In addition, larger distances were revealed at pre-fermentation and salted phetze stages, indicating greater changes of tastes occurred rather than those at post-fermentation stage.These results are in accordance with the prior implication by PCA (Fig.4a), which indicated that metabolites had remarkable effects on red sufu tastes.Changes of tastes at post-fermentation stage were illustrated in the score plot(Fig.4d), which was generated by combining PC1 (48.1%)with PC2 (27.2%).Samples (S7–S9) at the early stage of post-fermentation had larger distances, indicating the greater taste differences existed.However, samples at the late stage (S10–S12) located in the positive PC1 axis, were also clearly separated.Hence, combined with PCA, the actual performance of taste sensing system for the determination of red sufu maturity extent can be assessed.It could be considered as a complementary or alternative technique to the traditional analytical reference methods.

2.4 HCA and correlation analysis in terms of metabolits and taste atttributes

HCA is a multivariate analysis technique to classify samples into groups by measuring either distances or similarity among samples[36].It provides an overview of the changes of metabolites measured in twelve samples during fermentation (Fig.5a).The result demonstrated that samples were apparently clustered at the first level of cluster analysis,including group 1 (S1–S7) and group 2 (S8–S12).Samples at the latter period of post-fermentation stage (S10–S12)were clustered, which means that the quality of red sufu was getting stable at the end of fermentation.At the last cluster analysis, sample S12 was slightly different from S10 and S11, which presented that metabolites could be an effective method for the determination of sufu maturity extent.According to Fig.5a, significant differences in metabolites were clearly demonstrated, in terms of different fermentation times.Ala and Cys were strongly positively correlated with S3, while Pro, oxalic acid and formic acid were positively correlated with S6.These compounds were found to be the metabolites mainly contributing to the differentiation of red sufu at the end of pre-fermentation and salted phetze stages.During post-fermentation, lactic acid, acetic acid, Arg,Asp, and Ser were significantly responsible for changes at different times of post-fermentation.Among them, Asp and Ser had the most significant increase in S12, which could be considered as the potential indicators for the determination of red sufu maturity extent.Overall, HCA analysis suggested that each fermentation stage could have a positive influence on chemical compositions of red sufu and specific metabolites could be regarded as potential biomarkers to indicate fermentation progress.

Fig.5 HCA analysis (a) of detected metabolites, correlation matrix (b)in terms of taste attributes and correlation analysis (c) between taste attributes and metabolites in red sufu during fermentation

In order to evaluate the independence of taste attributes,a correlation matrix was carried out to determine correlations among taste attributes using Pearson’s correlation tests (Fig.5b).Previous studies have investigated the interactions between different taste attributes.Bitterness amino acids,including Phe and Tyr, are identified as umami taste enhancers in MSG/NaCl mixture at subthreshold levels[37].NaCl could mask bitterness taste and enhance umami taste[12].Astringency was negatively correlated with sourness,and sourness was decreased in the presence of NaCl and bitterness, as well as bitterness suppressed by NaCl and sourness[38].In addition, it has been identified that umami taste could contribute to the richness in beef consommé[39].According to Fig.5b, the correlation matrix showed that umami, richness, and saltiness, the major taste attributes in red sufu, were highly positive with each other.In addition,bitterness was positively associated with aftertaste-B and aftertaste-A, as well as negatively correlated with saltiness.Moreover, aftertaste-A was negatively correlated with umami, richness and saltiness.The interaction between taste attributes could provide a theoretical foundation for the taste development of red sufu products.

To illustrate the correlation between taste attributes and metabolites, the correlation analysis based on the calculation of the Pearson correlation coefficient andPvalue was shown in Fig.5c.Umami had a significantly positive correlation with Glu, His, succinic acid, NaCl and AAN (absolute values of Pearson’s correlation coefficient > 0.7,P< 0.05).Saltiness and astringency were positively correlated with NaCl, whereas the bitterness attribute was associated with Cys, Gly, and Ala.Besides, richness attribute was positively correlated with Asp, Met, acetic acid, lactic acid and reducing sugar.Aftertaste-A, as the unpleasant taste, was negatively correlated with Glu, Ala, Gly, Cys, Phe, and Val.The above results indicated that these compounds played an important role in the taste characteristics of red sufu.

3 Conclusion

A total of 17 FAAs and 6 organic acids, which had great effects on the taste of red sufu, were determined by IC and amino acid analyzer.Total FAAs and organic acids had increase significantly at pre-fermentation and postfermentation stages (P< 0.05).Among them, the levels of most FAAs and organic acids increased significantly in pre-fermentation and post-fermentation stages(P< 0.05).Statistical analysis methods, PCA and HCA,which successfully clustered red sufu samples with different fermentation stages into several groups, were used to characterize taste quality of red sufu.The results demonstrated that taste components of red sufu samples in different fermentation stages had great differences.Meanwhile, the overall taste profiles were analyzed by electronic tongue.Results showed that 8 taste attributes were identified, umami, saltiness, bitterness and richness were main tastes of red sufu, and taste attributes that aversive to most consumers, decreased at the last two months of postfermentation stage.The correlation analysis results shown that umami, richness, and saltiness were highly positive correlation with each other, and taste attributes of red sufu were positively correlated with physiochemical indexes(AAN, NaCl and reducing sugar), organic acid (lactic acid,acetic acid and succinic acid) and free amino acid (Glu, Asp,Cys, His and Met), negatively correlated with Ala, Gly, Cys,Phe and Val.

The statistical analysis of electronic tongue determination had similar results with chemical compositions analysis, indicating that both methods had potential application in evaluating taste quality, differentiating red sufu samples at different fermentation stages, and supervising quality of red sufu.Furthermore, the established discriminant model could evaluate red sufu maturity extent in the manufacturing process, and the results indicated that taste sensing system could be a rapid tool for taste evaluation of red sufu.Therefore, this research provides experimental evidence and scientific basis for evaluating the taste quality of red sufu.