Food Sci. Technol. Res., 18 (5), 747–754, 2012 Note

Changes in the Physicochemical and Volatile Flavor Characteristics of

Scomberomorus niphonius during Chilled and Frozen Storage

* Bin Zhang, Shang-Gui Deng and Hui-Min Lin

Department of Aquatic Product Processing and Storage, College of Food and Pharmacy, Zhejiang Ocean University, Zhoushan 316000, China

Received April 8, 2012; Accepted June 1, 2012

Changes in chemical, textural, and volatile flavor properties were investigated for fish ( niphonius) stored in cold rooms (5℃ and 0℃) and freezers (−20℃ and −55℃). Corre- lation and multivariate analysis showed a significant time-dependent relationship between total volatile base nitrogen (TVBN)/ trimethylamine (TMA) (Y) and storage time (X) for fish stored in cold rooms, with R = 0.996 − 0.997 values of Gompertz model (Y = a*exp(−exp(b − cX)), and there was a good linear rela- tionship between TVBN and TMA. Combined with the textural properties, the polynomial fitting model (Y = a + bX + cX2 +…, R = 0.982 − 0.991) was applied and elucidated the correlation between the hard- ness/springiness (Y) and TVBN (X), the rational function model (Y = (a + bX)/(1 + cX + dX2), R = 0.975 − 0.979) used for the chewiness (Y) and TVBN (X). The electronic nose analysis revealed that the variation of muscle volatile flavor compounds was found out along the PC1 to the right, and then along the PC2 to the upward and further to the downward based on the principal component analysis (PCA). Further- more, the linear discriminant analysis (LDA) had better distinction effect for the changes of fish flavor than PCA. Results from this study suggested that the texture analysis in combination with electronic nose techniques might be utilized as a rapid expeditious process for predicting quality and shelf life of the fresh fish or other aquatic products.

Keywords: Scomberomorus niphonius, TVBN, TMA, texture properties, electronic nose

Introduction specific spoilage organisms from various sources (Tzikas et The pelagic mackerel fishes are important fishery re- al., 2007; Shi et al., 2008). sources which are wildly distributed in all temperate and Freshness is the most common and fundamental criterion tropical seas, such as the East China Sea, Pacific coast of for the quality and acceptability of fish or fishery products. Japan, and regions of Indian and Russia (Bae et al., 2011). Microbiological, biochemical, and sensory analysis have The fish constitutes desirable components of a healthy diet, been used to evaluate the freshness and quality of fish during including polyunsaturated fatty acids, essential amino acids, handling and storage (Ozogul et al., 2011). The biochemical vitamins, and minerals. However, this species is highly per- properties, such as total volatile base nitrogen (TVBN) and ishable during processing and storage, due to large amounts trimethylamine (TMA), have been frequently used as the in- of non-protein nitrogen, high content of fat and autolytic dicators for the quality of fish, due to the correlation between enzymes in tissues, which deteriorates rapidly post-mortem organoleptic characteristics and microbial spoilage. How- resulting in formation of an obvious off-taste and soften tex- ever, with respect to the degree of spoilage or microbial me- ture (Simeonidou et al., 1998). What’s more, these problems tabolites, the biochemical analyses are tedious or time-con- will be accelerated by storage conditions such as abominable suming and may encounter serious safety concerns through temperature and moisture in environment and overgrowth of the contact of chemical reagents, such as toluene and trichlo- roacetic acid used in TVBN and TMA determination (Malle *To whom correspondence should be addressed. et al., 1987). The sensory evaluation is the most popular way E-mail: [email protected] of assessing quality of fish, which provides the fast, simple 748 B. Zhang et al. and visible quality information for the consumers (Goulas et homogenate filtrate was measured using a digital pH-meter al., 2007), but it is often subjective and inaccurate, therefore, (PHS-25, Shanghai Precision & Scientific Instrument Co. requires trained panels to perform analysis. In addition, the Ltd., Shanghai, China). sensory evaluations also may introduce serious safety risk Biochemical analysis The total volatile base nitrogen factors as those of toxin production (such as biogenic amines (TVBN) was determined by steam-distillation of trichloro- in mackerel fish). Texture of raw fish can be measured by acetic acid (TCA)-fish extract, using the modified method of different methods (puncture, compression, shear and tensile Malle et al. (1987). The fish muscle (10.0 g) was added to stress). Among them, the shearing force and compression 100 mL of 0.60 mol/L TCA solution and homogenized at a methods are recommended for use with fresh fish (Martinez speed of 10,000 rpm for 2.0 min (0 − 4℃). The homogenate et al., 2007). Some of the information is available about the was centrifuged at 5000 g for 5 min (LD5-2A Centrifuge, texture assessment of mackerel fish, but little is known about Beijing Lab Centrifuge Co., Ltd., Beijing, China). Five mil- the correlation among the textural characteristics, spoilage liliters of supernatant were transferred to a Kjeldahl-type dis- degree and storage behavior (Mishima et al., 2005). tillation tube followed by additions of 5.0 mL of 0.80 mol/L It is necessary to use the chemical and physical methods NaOH solution, a drop of phenolphthalein indicator and two when the purpose is to adequately define and assess the fish drops of silicon oil. Steam-distillation was pipetted into the quality, and ensure that those results show good agreement inner ring for 6.0 − 7.0 min. The distillate was collected in a with the objective tests. Thus, the objectives of this study flask containing 10.0 mL of 0.50 mol/L boric acid solution. were to investigate changes in the quality of mackerel fish Finally, the boric acid distillate was titrated with 0.01 mol/L during storage at different temperatures over time based on HCl solution until the color turned to bluish violet. biochemical analysis and establish a model for rapid assess- The trimethylamine (TMA) was determined according to ment of quality and shelf life using textural and volatile fla- the procedure of AOAC with minor modifications (AOAC vor characteristics obtained from this study. 1972). Ten grams of fish muscle were blended with 70 mL of distilled water and then homogenized at a speed of 10,000 Materials and Methods rpm for 1.0 min. The homogenate was mixed with 10.0 mL Samples preparation and storage Fresh mackerel of 40% (w/v) TCA solution, vibrated for 5.0 min and filtered (Scomberomorus niphonius) of 30 − 35 cm in body length through a Whatman 1.0 filter paper (0 − 4℃). The filtrate with 400 − 500 g in body weight were bought at the Nan- (5.0 mL), referred as sample extract, was placed in a 25-mL zhen fishing port in Zhoushan city, China. The fish samples colorimetric tube. To the sample extract, 10.0 mL of 10% (about 200 fish, be stiff in) were kept in ice and transported (v/v) formaldehyde solution, 10.0 mL of anhydrous toluene immediately to the laboratory. The following preparation and 10.0 mL of 7.2 mol/L potassium carbonate solution were process was carried out in a cold room at a temperature of 0 added. The mixture was shaken for approximately 1.0 min, − 4℃. The fresh fish (K-value about 7.0%) were carefully and left to stand for another 20.0 min. Five milliliters from packed in polyethylene bags and divided into four treatment the toluene layer were transferred into a dry tube contain- batches for storage at 5, 0, −20 and −55℃, respectively. ing 5.0 mL of 0.02% (w/v) picric acid toluene solution and Before being analyzed, frozen samples (−20℃ and −55℃) measured using a spectrophotometer (UV 2102 PC, UNICO, were thawed overnight at refrigerator temperature (4℃). The Shanghai, China) at 410 nm. samples stored at 5 and 0℃ were randomly taken and ana- Texture profile analysis Texture profile analysis (TPA) lyzed every day, while the samples stored at −20 and −55℃ of the fish was measured by using a texture analyzer (TMS- were randomly chosen and analyzed every ten days. The PRO, FTC, Sterling Virginia, USA). The sample was place fish were headed and filleted manually using a sterile scalp on the instrument’s platform and a P/50 cylindrical Perspex and forceps. Specimens of dorsal ordinary muscles (taken probe (50 mm diameter) was selected to simulate chewing from three fish) with lapse of time were used for pH, TVBN, process. TPA was performed with the following conditions: TMA and volatile flavor measurements. Each property was constant test speed 1.0 mm/s, sample deformation 30%, hold conducted at least three times in a randomized design and all time between cycles 3.0 s, and trigger force 0.05 N. Tex- measurements were carried out at least twice. ture analysis parameters were calculated using FTC-PRO pH measurement The pH value of fish muscle was de- software (FTC, Sterling Virginia, USA) from the force-time termined according to the method of Benjakul et al. (1997). curves generated from each sample as described by Bourne Ten grams of each sample were blended with 40 mL of (1978). distilled water in a blender (T18 ULTRA–TURRAX, IKA, Electronic nose analysis Electronic nose determination Staufen, Germany) for 60 s at 0 − 4℃. The pH value of was performed with a PEN3 instrument (Airense, Schwerin, Quality Assessment of Mackerel Fish 749

Germany). Processed fish muscle samples were cut into pH of fresh mackerel had a value of 6.12, which increased uniform pieces of approximately 5.0 g, and filled into a 10- significantly to 7.11 in samples stored at 5℃ on the 6th day mL vial which was kept at room temperature (25 ± 2℃) and to 6.96 in samples stores at 0℃ on the 7th day. And both for 20 min to allow equilibration before static headspace samples also showed a spoilage sign with obvious off-odor. sampling. The following measurement conditions were as This increment was attributed to the decomposition of nitrog- follows: sample interval 1.0 s, pre-sampling time 5.0 s, zero enous compounds with accumulation of alkaline compounds point time 10.0 s, measurement time 60.0 s, flush time 60.0 such as ammonia compounds, dimethylamine, TMA and s, chamber flow 400 mL/min, initial injection flow 400 mL/ other biogenic amines mainly derived from microbial growth min, and automatic dilution program. For the evaluation of or activity of endogenous enzymes (Simeonidou et al., 1998; electronic nose sensor signal response of the sensor array, Sallam et al., 2007). However, such increment was incon- principal component analysis (PCA) was performed, which spicuous in frozen samples (−20 and −55℃) during the 70 displays a two or three dimensional view of the training data days of storage. This was probably due to the concentrations in the pattern file. Moreover, linear discriminant analysis of glycogen immediately post-mortem and subsequently the (LDA) was also applied in the evaluation, which uses the levels of generated lactic acid maintaining the low initial pH information of class membership in order to optimize the of muscle (Tzikas et al., 2007), or inhibition on the liberation resolution between classes (Poulli et al., 2005). of inorganic phosphate and ammonia caused by the enzy- Statistical analyses Measurements of each property of matic degradation of ATP (Riebroy et al., 2007), which sug- the sample were conducted at least three times in a random- gested the quality of mackerel was remarkably persevered by ized design and all measurements were performed twice. the refrozen storage. Statistical analysis and significant differences were conduct- TVBN and TMA contents TVBN and TMA are consid- ed using the SPSS package (SPSS 13.0 for windows, SPSS ered as valuable tools in evaluating fish quality. Schormuller Inc., Chicago, IL, USA) and Origin 8.1 software (OriginLab, (1968) and Ludorff et al. (1973) proposed a TVBN value Northampton, USA). The appropriate curve fitting and mod- of 35 mg/100 g of fish muscle as the acceptable limit for eling analysis were fitted by using CurveExpert 1.40 program edible flesh. A similar limit of TVBN value of 30 mg/100 (Copyright© 2010 Daniel G. Hyams), which employs a large g for fresh marine fish (green dotted line in Fig.1 (a)) was number of regression models (including linear, quadratic, or- established in China (i). In the study, the initial TVBN level der polynomial, exponential family, power law family, yield- in mackerel fish was 2.7 mg/100 g in accordance with the density models, and so on) to represent the data in the most previous report (Mbarki et al., 2009), indicating a good qual- precise and convenient way. ity of the samples (Fig. 1). Subsequently, TVBN values of samples stored at 5 and 0℃ increased to 34.5 mg/100 g on Results and Discussion the 5th day and to 34.2 mg/100 g on the 7th day, respectively, pH variation The pH value, around 6.0 ~ 6.5 for fresh and the samples became slightly slimy with dull eyes and a fish post-mortem, will increase due to bacterial growth and weak fishy odor. However, the refrozen samples were all in biochemical reactions, the limit of which for fish acceptabil- good condition and the TVBN values remained lower than ity is usually 6.8 ~ 7.0 (Metin et al., 2001). In this study, the 10 mg/100 g even after 70 days of storage. These signifi-

Fig. 1. TVBN (a) and TMA (b) values evolution in the mackerel fish flesh during cold (5℃ and 0℃) and frozen (−20℃ and −55℃) storage. Vertical bars denote standard deviation. 750 B. Zhang et al. cantly lower TVBN levels were attributed to the inhibition tween TVBN (Y) and storage time (X) under specified con- of growth of microorganisms and the muscle autolytic reac- ditions (5 and 0℃, 0 ~ 7 d) by using CurveExpert 1.40 pro- tions. gram (Copyright© 2010 Daniel G. Hyams). Results showed TMA is also often documented as a biochemical index to that there was a good time/temperature-dependent relation- assess quality and is the main responsible for the unpleasant ship between the two characters, with correlation coefficient fishy odour. Since growth of bacteria is usually prohibited at 0.996 value of Gompertz model (Y = a*exp(−exp(b − cX)), −18℃, the increase in TMA may be formed from enzymatic a = 44.571, b = 1.194, c = 0.501) for samples stored at 5℃ degradation of trimethylamine oxide (TMAO) (Simeonidou and correlation coefficient 0.997 value of Gompertz model (a et al., 1997). The TMA limit of fish acceptable for human = 218.988, b = 1.150, c = 0.128) for samples stored at 0℃. consumption was set as 10 mg/100 g fish (green dotted line Moreover, the correlation analysis result indicated that, the in Fig.1 (b)). The initial TMA content of samples used in this TVBN values (Y) had good linear relation with TMA values study was 1.1 mg/100 g, which increased in all samples dur- (X), with correlation coefficient 0.988 for samples stored at ing the storage period. Significantly higher (p < 0.05) values 5℃ (Y = 2.481 + 2.772X, 0 ~ 7 d) and correlation coefficient of TMA were observed in cold-storage samples than in fro- 0.988 for samples stored at 0℃ (Y = −1.168 + 3.285X, 0 ~ 7 zen-storage samples. TMA contents increased to 11.2 mg/100 d). Results from the above curve-fitting models, TVBN/TMA g for samples at 5℃ on the 5th day and to 10.2 mg/100 g for values of mackerel fish could be quickly speculated accord- samples at 0℃ on the 7th day both exceeded the acceptable ing to its storage temperature and time, further to assess the limit. However, the TMA values if samples stored at −20 and fish storage quality and predict the shelf life of the storage −55℃ remained below 5.0 mg/100 g throughout the entire products. storage period. And the values of TVBN and TMA at −20 Textural properties Hardness is the measurement of and −55℃ increased slightly especially after 40 days stor- how resistant a solid matter is to various kinds of perma- age, which might be attributed to the dryness and drip loss nent shape change when a force is applied. In the study, the of fish muscle at lower temperature values (Riebroy et al., obvious variation was that the hardness of mackerel fish 2007). decreased dramatically with a significant loss of 54.7% for Correlation and multivariate analysis was carried out be- samples at 5℃ and 34.5% for samples at 0℃ after 7 days

Fig. 2. Hardness (a), springiness (b), cohesiveness (c) and chewines (d) values evolution in the mackerel fish during cold (5℃ and 0℃) and frozen (−20℃ and −55℃) storage. Vertical bars denote standard deviation. Quality Assessment of Mackerel Fish 751 of storage (Fig. 2). However, the frozen-storage samples To compare the biochemical indices of mackerel fish showed significant differences in the hardness variables, and with its texture properties, the polynomial fitting model was only lose 17.3% for samples at −20℃ and 4.8% for samples applied and elucidated the correlation between the hardness at −55℃ after 70 days of storage. Furthermore, the hard- (Y) and TVBN (X), with R = 0.984 for 5℃-storage samples 2 ness values of samples at −55℃ were lower than at −20℃ (Y = a + bX + cX , a = 47.637, b = 0.549, c = −0.029) and 2 during all the storage. This was attributed to larger ice crys- R = 0.982 for 0℃-storage samples (Y = a + bX + cX , a = tal formed at −20℃, which produced stronger mechanical 51.078, b = −0.210, c = −0.009) over 7 days of storage. The damage to the muscle tissue and separation of myofibrillar polynomial fitting model was also applied for the correla- proteins, making the muscle softer and less elastic (Dai et tion between springiness and TVBN (R = 0.983 − 0.991), al., 2008). The undesirable hardness changes might also be cohesiveness and TVBN (R = 0.976 − 0.988) (models were caused by the activity of autolytic enzymes (e.g. collagenase, not shown). Moreover, the most appropriate fit between the ATPase) hydrolyzing proteins and breaking down the con- chewiness values (Y) and TVBN (X) was in accordance with nective tissues, further to allow rapid multiplication of spoil- the rational function model. The correlation coefficient were age microorganisms and promote the progress of spoilage obtained as 0.975 for 5℃-storage samples (Y = (a + bX)/(1 (Hattula et al., 2001). + cX + dX2), a = 11.549, b = −0.239, c = −0.042, d = 0.001) Springiness indicates the elasticity of muscle that can be and 0.979 for 0℃-storage samples (Y = (a + bX)/(1 + cX + stretched and returned to its original length. Cohesiveness in- dX2), a = 12.018, b = −0.854, c = −1.125, d = 0.002), respec- dicates the property of being cohesive and sticky of the fish. tively. These established models gave a reasonable indication From the testing results, the springiness and cohesiveness of freshness in terms of quality parameters, which might be parameters decreased significantly with the time of storage, referred to the quick evaluation of the raw seafood. although two different change stages were observed in all Volatile flavor evaluation To investigate whether the storage samples. The first stage was from 0 ~ 2 d for cold- electronic nose was able to distinguish different quality de- storage samples (0 ~ 30 d for refrozen-storage samples) with gree of storage fish, PCA and LDA were applied to train the relatively smooth and steady changes. The second stage was collected volatile flavor produced by fish samples. PCA re- from 3 ~ 7 d (30 ~ 70 d for refrozen-storage samples), in sults were showed in Fig. 3 to describe the variation among which significant changes occurred over the time. Hattula et samples stored at different temperatures. The first principal al. (2001) reported that the fish spoilage caused by the activ- component (axis-x, PC1) explained 93.15%, and the second ity of autolytic enzymes in tissue appeared to have a major principal component (axis-y, PC2) explains 1.67% of the effect on texture deterioration. total variance with value 94.82% for 5℃-storage samples. Chewiness is the mouth feel sensation of labored mas- Through examining scores plotted in the area defined by tication due to sustained, elastic resistance from the fish, the two principal components, the relative clear separa- defined as the product of hardness × cohesiveness × springi- tion of volatile flavor was found among the various clusters ness, which had a similar changing trend to that of springi- representing quality degree of fish muscle. Besides, each of ness and cohesiveness (Fig. 2). A similar report showed that volatile flavors collected from different samples had a strong these lower textural properties of samples stored for a long convergence and was distinguished with each other. For the time were in accordance with the decrease in water hold- PCA results of samples stored at 0℃, the total variance value ing capacity (Riebroy et al., 2007). From the results, storing of 83.86% showed a shift erratic variation at different storage at temperatures of −20 and −55℃ delayed post- state. There was no particular trend during the storage time, mortem deterioration of quality, and the decrease in textural particularly day 0 − 2 overlapping heavily. It was perhaps be- properties perhaps was attributed to the water retaining cause the quality variation degree of the samples were limit- capacity and drip loss of fish muscle at lower temperature ed at the first 0 − 2 day, thus the difference in volatile among values, leading to the ordered formation of muscle protein day 0, 1 and 2 samples were not obvious, causing the over- structure (Riebroy et al., 2007). More specifically, the values lapped phenomena on PCA results. What’s more, the PCA of hardness, springiness, cohesiveness, and chewiness of the results suggested that the variation of muscle volatile flavor samples in the second stage had a tendency to decrease at compounds was found out along the PC1 to the right, and all temperature, the reasons might be the breakdown of the then along the PC2 to the upward and further to the down- myofibrillar structure, be ascribed mainly to the activity of ward. For the results of samples stored at −20 or −55℃, all lysosomal cathepsins, in particular cathepsin B, cathepsin L the volatile flavors collected from different storage-time had and cathepsin D (Godiksen et al., 2009), which were highly a strong convergence and could not distinguished with each active in the 5 and 0℃ storage fish. other, which indicated there were no obvious changes in the 752 B. Zhang et al.

Fig. 3. PCA analysis for mackerel fish during cold (5℃ and 0℃) storage.

Fig. 4. LDA analysis for mackerel fish during cold (5℃ and 0℃) storage. flavor appearance for the groups during the lower tempera- eigenvalue (Zhou et al., 2011). ture storage. Furthermore, the obvious fluctuant cluster in the PCA After analysis of the same data set using LDA, the 0 − 7 plot occurred on the 4th and the 7th days, representing the d storage samples were also distinguishable from each other sample stored at 5 and 0℃ respectively. It was suggested (Fig. 4, the data of −20 and −55℃ storage samples were not that the massive changes occurred in the volatile flavor of shown because there were no obvious changes in the ap- the samples, which was perhaps a weak fishy spoilage. These pearances). In the 5℃-storage plot, LDA function 1 (axis-x, results were precisely in conformity with the determination LD1) and function 2 (axis-y, LD2) accounted for 91.14% and results of TVBN, TMA, and the texture properties. In the 5.72% respectively, with the total variance value of 96.86%. LDA plot, the abnormal ellipse occurred on the 5th and the This analytic approach was efficient to separate the differ- 7th days, representing the sample stored at 5 and 0℃ respec- ent quality degree of sample from others. In the 0℃-storage tively, the trend of which was basically consistent to the PCA plot, the samples were also distinguishable from each other results. Taken together, both PCA and LDA had good distinc- obviously (the total variance value of 93.58%). These results tion for the different fish samples. Moreover, the LDA ap- indicated that the distinguishable performance for quality by proach had better distinguishable result than PCA approach. LDA approach were better than PCA analysis. This could be Combined with the biochemical and textural analysis results, accounted for the facts that LDA summarized the separation it was possible to use electronic nose technique to predict of samples among groups into a reduced space, while PCA the quality characteristics and the remaining shelf life of the was a projection method transforming the original variables storage fish. onto new ones, orthogonal and arranged according to their Quality Assessment of Mackerel Fish 753

Conclusion Goulas, A.E. and Kontominas, M.G. (2007). Effect of modified The biochemical and physical characteristics were deter- atmosphere packaging and vacuum packaging on the shelf-life mined to assess the quality and acceptability of mackerel fish of refrigerated ( japonicus): biochemical during storage at various temperatures. There were high de- and sensory attributes. Eur. Food Res. Technol., 224, 545-553. grees of agreement between TVBN, TMA, texture properties Ludorff, A. and Meyer U. (1973). Fische und Fisherzeugnisse. Paul and volatile flavor measurements. TVBN and TMA increased Parey Verlag. Berlin Hamburg, pp. 219-309. in all samples and showed a good time-dependent relation- Malle, P. and Tao, S.H. (1987). Rapid quantitative determination of ship simulated by classical Gompertz model. The hardness, trimethylamine using steam distillation. J. Food Prot., 50, 756- springiness, cohesiveness and chewiness data indicated that 760. the texture properties of the fish had a reduction trend in all Martinez, O., Salmerón, J., Guillén, M.D. and Casas, C. (2007). variables after a long time storage, which could be fitted by Textural and physicochemical changes in salmon (Salmo salar) polynomial or rational function model. The volatile flavor treated with commercial liquid smoke flavourings. Food Chem., evaluation showed the electronic nose could be used to dis- 100, 498-503. criminate the fish quality and acceptability. Taken together, Mbarki, M., Sadok, S. and Barkallah, I. (2009). Quality changes these results demonstrated the usefulness of TPA and volatile of the Mediterranean horse mackerel ( mediterraneus) flavor analysis in providing quantitative parameters for as- during chilled storage: The effect of low-dose gamma irradiation. sessing quality and shelf-life of fish. Radiat. Phys. Chem., 78, 288-292. Metin, S., Erkan, N., Varlik, C. and Aran, N. (2001). Extension of Acknowledgment We thank the Program of International S&T shelf-life of chub mackerel (Scomber japonicus Houttuyn 1780) Cooperation (2010DFB34220 and 2012DFA30600), the National treated with lactic acid. Eur. Food Res. Technol., 213, 174-177. Natural Science Fund project of China (31071628 and 31201452), Mishima, T., Nonaka, T., Okamoto, A., Tsuchimoto, M., Ishiya, T., and the Science Fund of Zhejiang Ocean University (21135011010 Tachibana, K. and Tsuchimoto, M. (2005). Influence of storage and 21135003411). Special thanks to Dr. Yi-Cheng SU from Or- temperatures and killing procedures on post-mortem changes in egon State University. the muscle of horse mackerel caught near Nagasaki Prefecture, Japan. Fisheries Sci., 71, 187-194. References Ozogul, Y. and Balikci, E. (2011). Effect of various processing AOAC. (1978). Official Method of Analysis of AOAC Inl. 35th ed. methods on quality of mackerel (Scomber scombrus). Food Bio- Method 971.14. Association of Official Analytical Communities, process Technol., 29, 1-8. Gaithersburg, MD, USA. Poulli, K.I., Mousdis, G.A. and Georgiou, C.A. (2005). Classifica- Bae, J.H., Yoon, S.H. and Lim, S.Y. (2011). Heavy metal contents tion of edible and lampante virgin olive oil based on synchronous and chemical compositions of Atlantic (Scomber scombrus), Blue fluorescence and total luminescence spectroscopy. Anal. Chim. (Scomber australasicus), and Chub (Scomber japonicus) mack- Acta., 542, 151-156. erel muscles. Food Sci. Biotechnol., 20, 709-714. Riebroy, S., Benjakul, S., Visessanguan, W. and Tanaka, M. (2007). Benjakul, S., Seymour, T.A., Morrissey, M.T. and An, H. (1997). Effect of iced storage of bigeye snapper (Priacanthus tayenus) on Physicochemical changes in Pacific whiting muscle proteins dur- the chemical composition, properties and acceptability of Som- ing iced storage. J. Food Sci., 62, 729-733. fug, a fermented Thai fish mince. Food Chem., 102, 270-280. Bourne, M.C. (1978). Texture profile analysis. Food Technol., 32, Sallam, K.I., Ahmed, A.M., Elgazzar, M.M. and Eldaly, E.A. (2007). 62-66. Chemical quality and sensory attributes of marinated Pacific Dai, Z.Y., Cui, Y.N. and Wang, H.H. (2008). Changes of textural saury (Cololabis saira) during vacuum-packaged storage at 4℃. properties of cultured Pseudosciaena Crocea muscle under dif- Food Chem., 102, 1061-1070. ferent frozen storage conditions. Food Fermentation Ind., 34, Schormuller J. (1968). Handbuch der Lebensmittel Chemie. Band 188-191. III/2 Teil. Trierische Lebensmittel Eier, Fleisch, Buttermilch. Hattula, T., Elfving, K., Mroueh, U.M. and Luoma, T. (2001). Use Springer-Verlag, New York, pp. 1493-1497. of liquid smoke flavourings as an alternative to traditional flue Shi, Q.L., Xue, C.H., Zhao, Y., Li, Z.J., Wang, X.Y. and LUAN, D.L. gas smoking of rainbow trout fillets (Oncorhynchus mykiss). (2008). Optimization of processing parameters of horse mackerel LWT-Food Sci. Technol., 34, 521-525. (Trachurus japonicus) dried in a heat pump dehumidifier using Godiksen, H., Morzel, M., Hyldig, G. and Jessen, F. (2009). Contri- response surface methodology. J. Food Eng., 87, 74-81. bution of cathepsins B, L and D to muscle protein profiles corre- Simeonidou, S., Govaris, A. and Vareltzis K. (1997). Effect of fro- lated with texture in rainbow trout (Oncorhynchus mykiss). Food zen storage on the quality of whole fish and flllets of horse mack- Chem., 113, 889-896. erel (Trachurus trachurus) and mediterranean hake (Merluccius 754 B. Zhang et al.

mediterraneus). Z. Lebensm. Unters. F. A., 204, 405-410. ing storage in ice. Food Control, 18, 1172-1179. Simeonidou, S., Govaris, A. and Vareltzis, K. (1998). Quality as- Zhou, B. and Wang, J. (2011). Detection of insect infestations in sessment of seven Mediterranean fish species during storage on paddy field using an electronic nose. Int. J. Agr. Biolog., 13, 707- ice. Food Res. Int., 30, 479-484. 712. Tzikas, Z., Ambrosiadis, I., Soultos, N. and Georgakisa, S. (2007). Quality assessment of Mediterranean horse mackerel (Trachurus URL Cited mediterraneus) and blue jack mackerel (Trachurus icturatus) dur- i) http://www.aqsiq.gov.cn/ (Apr.2, 2008).