Clustering of commercial fish sauce products based on an e-panel technique

Fish sauce is a brownish liquid seasoning with a characteristic flavor that is produced in Asian countries and limited areas of Europe. The types of fish and shellfish and fermentation process used in its production depend on the region from which it derives. Variations in ingredients and fermentation procedures yield end products with different smells, tastes, and colors. For this data article, we employed an electronic panel (e-panel) technique including an electronic nose (e-nose), electronic tongue (e-tongue), and electronic eye (e-eye), in which smell, taste, and color are evaluated by sensors instead of the human nose, tongue, and eye to avoid subjective error. The presented data comprise clustering of 46 commercially available fish sauce products based separate e-nose, e-tongue, and e-eye test results. Sensory intensity data from the e-nose, e-tongue, and e-eye were separately classified by cluster analysis and are shown in dendrograms. The hierarchical cluster analysis indicates major three groups on e-nose and e-tongue data, and major four groups on e-eye data.

cluster analysis indicates major three groups on e-nose and e-tongue data, and major four groups on e-eye data.
& 2018 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Subject area
Food science More specific subject area Food quality assessment using electronic sensor technology Type of data Table and  The fish sauce products used were made by various companies in Japan, Thailand, Vietnam, China, the Philippines, and Italy. Experimental features The smells, tastes, and colors of commercially available fish sauce products were assessed by e-nose, e-tongue, and e-eye analysis. Data source location Tokyo, Fukuoka, and Abashiri, Japan Data accessibility Data are presented in this article

Value of the data
The availability of these data will enable a discussion of the influence of factors such as raw materials and the production process on smell, taste, and color development during fish sauce product manufacture.
The data presented will allow estimation of fish sauce product consumer preferences in different countries.
The smell, taste, and color data provided will be useful for culinary applications and the development of fish sauce products.

Data
Fish sauce is produced from seafood of various types. Table 1 displays the geographic origins and ingredients of the commercially available fish sauce products analyzed. Anchovy and sardine appear to be widely used in fish sauce production in Thailand and featured in fish sauce products from Vietnam, China, and Italy. A fish sauce product from the Philippines contained mackerel. Flying fish, which is often used as dashi, a broth employed in various dishes in Japan (especially in the Kyushu area), was found in several Japanese fish sauce products. Bonito is also employed for dashi preparation throughout Japan, and was also present in many of the fish sauce products tested from this country. In addition, Japanese sandfish, tuna, cod, sea bream, cutlassfish, deep-sea smelt, sea urchin, oyster, shrimp, and squid were found to be used in the manufacture of fish sauce products in Japan. Commercial fish sauce products from Japan, Thailand, Vietnam, China, the Philippines, and Italy clustered into three major groups when the e-nose data were analyzed (Fig. 1). Cluster analysis of the e-tongue data also identified three groups (Fig. 2), whereas that of the e-eye data yielded four major groups (Fig. 3).

Design
Fish sauce possesses distinctive smells, tastes and colors depending on the ingredient fish, fermentation process, geographic origins. Here we analyzed these features of fish sauces from Asian countries as well as Italian products. To avoid the effects of subjective assessment, we used an e-nose, e-tongue, and e-eye to precisely analyze the smell, taste, and color of fish sauce products from several countries. The e-nose and e-tongue consist of arrays of non-selective gas and liquid sensors with broad and partially overlapping selectivity towards the compounds present in a sample [1]. In addition, an e-eye was used to distinguish color components using camera-equipped apparatus and computer-assisted analysis. Of the commercial fish sauce products selected, 30 were produced in Japan, 11 were from Thailand, two derived   V2  P1  C1  I1  J21  J26  J24  J25  J22  J23  T4  T1  T2  T3  T6  T5  T7  T8  T10  T9  T11  V1  J7  J11  J9  J8  J10  J12  J19  J14  J16  J13  J17  J20  J6  J4  J5  J15  J18  J1  J2  J3  J28  J29  J27  CON   from Vietnam, and one was from each of China, the Philippines, and Italy ( Table 1). The list of numerical data indicating signal intensities on e-nose, e-tongue and e-eye sensors was brought to hierarchical cluster analysis to classify the fish sauces based on their smell, taste and color.

Materials
All of the 46 fish sauce products were purchased at local markets in Tokyo, Fukuoka, and Abashiri (Japan). The ingredients of each product as listed on their labels are summarized in Table 1. For analysis, the products were assigned product IDs as follows: J1-J30 for the Japanese products; T1-T11 for the Thai products; V1 and V2 for the Vietnamese products; and P1, C1, and I1 for the Filipino, Chinese, and Italian products, respectively.

Evaluation of smell with e-nose analysis
E-nose analysis was performed with the αFOX 4000 smell analysis system (Alpha M.O.S., Toulouse, France), which has 18 metal oxide gas sensors for different selectivity patterns. Fish sauce product samples (0.5 g) were collected in 10-ml vials, placed in an autosampler, and analyzed under the following conditions: syringe temperature, 50°C; oven temperature, 40°C; injection speed, 2 ml/s. The analyzed data was obtained as numerical values of signal intensities.

Evaluation of taste with e-tongue analysis
E-tongue analysis was performed with the αASTREE taste analysis system (Alpha M.O.S.), which has seven liquid potentiometric sensors (SRS, GPS, STS, UMS, SPS, SWS, and BRS) and a reference electrode (Ag/AgCl). Fish sauce products diluted 21-fold were collected in a beaker and placed in an autosampler. Each sensor was immersed in the samples for 120 s at 20-25°C with agitation to elicit a sensor response. A stable response value was then recorded at 120 s. The analyzed data was obtained as numerical values of signal intensities.

Evaluation of color with e-eye analysis
E-eye analysis of fish sauce products was carried out using an IRIS VA300 visual analyzer (Alpha M.O.S.) with a charge-coupled device camera. Five milliliters of fish sauce product was collected in a transparent plastic dish and placed in the measurement chamber. The color of the surface of each sample was measured three times. The collected color data were represented by IRIS color codes, which encompass 4096 colors. The analyzed data was obtained as numerical values of signal intensities.

Statistical analysis
Hierarchical cluster analysis was used to classify the fish sauces based on the smells, tastes and colors analyzed by e-nose, e-tongue and e-eye sensor sets. The datasets, comprising a series of sensor values from the e-nose, e-tongue, and e-eye tests, were subjected to cluster analysis by Ward's method using the program R-3.4.2 (http://www.R-project.org) [2].