DSpace at VNU: FLAVOR OF COOKED CATFISH (PANGASIUS HYPOPHTHALMUS) FILLETS AS PREDICTION FROM RAW FRESH AND FROZEN-TO-THAWED

This study also aimed to examine sensory characteristics of raw catfish fillets fresh and frozen as predictors of sensory properties of cooked catfish fillets.. The lexicon, in addition,

FLAVOR OF COOKED CATFISH (PANGASIUS HYPOPHTHALMUS)

FILLETS AS PREDICTION FROM RAW FRESH AND

FROZEN-TO-THAWEDjoss_363 12 25

1 Institute of Biotechnology and Food Technology, Ho Chi Minh City University of Industry, Ho Chi Minh-City, Vietnam

2 Faculty of Chemical Engineering, Ho Chi Minh City University of Technology, Ho Chi Minh-City HCM, Vietnam

3 Corresponding author TEL: +84-908191277;

FAX: +84-8-38636984; EMAIL:

dzung@hcmut.edu.vn

Accepted for Publication November 9, 2011

doi:10.1111/j.1745-459X.2011.00363.x

ABSTRACT

This study aimed to develop a lexicon of sensory descriptors for Vietnamese catfish

(Pangasius hypophthalmus) fillet products, and to investigate the use of sensory

char-acteristics of raw and frozen fillets to predict the flavor of cooked catfish fillets Descriptive Analysis (DA) was applied by a panel of 11 trained panelists to samples of raw fresh fillets and frozen fillets at three stages: frozen, thawing and thawed Samples were also cooked at 200C for 25 min to evaluate their flavor A lexicon of sensory attributes was generated for three types of samples: raw fresh, frozen-to-thawed and cooked The validity of this lexicon has been discussed Raw fresh and frozen-to-thawed samples were shown to be predictors for sensory characteristics of cooked samples Sensory quality of catfish fillets was indicated by textural properties and flavors more than appearances and colors

PRACTICAL APPLICATIONS

This study provides Pangasius processors with sensory lexicons of Pangasius fillets to

support the development of their quality control plans The study also provides pro-ducers and purchasers with methodological guidelines for making prediction of

flavor of Pangasius products at different phases of the processing: raw fresh, frozen,

thawed and cooked

INTRODUCTION

According to the Pangasius Aquaculture Dialogue of World

Wildlife Fund in 2010, pangasius farming is the world fastest

growing type of aquaculture The production has increased

dramatically from approximately 400,000 tons in 2005 to

825,000 tons in 2006 and reached 1 million tons in 2007,

according to the Vietnam Association of Seafood Exporters

and Producers (World Wildlife Fund [WWF] 2010b)

Pan-gasius is primarily farmed in Asia, mostly in Vietnam whose

world market share was 89% in 2005 and whose exported

value was $737 million in 2007 (WWF 2010a) Vietnam,

however, is facing strong competition from many other

catfish producing countries, such as Indonesia, China and the

U.S.A According to the World Market and Trade report of

USDA (2008), U.S fish and seafood exports topped $4 billion,

and China remained the top fish and seafood exporter with

the exported value of more than $8.5 billion in 2007

There-fore, to be well competitive, Vietnamese pangasius products

need (1) to meet international quality standards such as Codex Alimentarius, Association of Official Agricultural Chemists, American Society for Testing and Materials (ASTM) or EU standards; and (2) to possess outstanding typical properties These goals could be achieved through an appropriate sensory quality program

Sensory evaluation has been often used in the fishing industry as a part of quality control management to ensure that the products will meet certain standards or consumers’

expectations (see Simeonidou et al 1997; Codex Alimenta-rius 2001; Pons-Sánchez-Cascado et al 2006) However, the

most concern of a sensory program in fishery is always the quality of freshness Methods have been developed specifi-cally for judging freshness by smell, color, appearance, taste and texture They include the EU scheme (Man and Jones

2000), Quality Index Method (QIM) (Martinsdóttir et al.

2001) and Torry scheme (Man and Jones 2000) These

schemes have been proven to have many advantages in

prac-tice, especially in making purchase decision However, these

systems are problematic in the sense that they give a single

numerical value to a broad range of characteristics QIM, for

instance, evaluates the freshness of fish using only one

indica-tor, the quality index, which is the sum of the scores for all

characteristics such as the eyes, skins and gills This system

groups attributes with major changes into one grade,

regard-less their difference in nature QIM scheme for cod, for

instance, groups “fresh, seaweedy, and metallic” with score of

0; “yeast, bread, beer, sour milk” with score of 2; “acetic acid,

sulfuric, very sour” of score 3 (Martinsdóttir et al 2001).

Grading systems, therefore, are only useful to deal with

quality standards but not appropriate for investigating

sensory typical characteristics of fishery products

Descriptive Analysis (ASTM International 1992; ISO

2003), on the other hand, is recommended as a good tool for

quality control and product development According to Stone

(1992), the method is based on the principle of a panelist’s

ability to verbalize perceptions of a product in a reliable

manner The best advantage of descriptive analysis is that it

allows an integral understanding of the sensory

characteris-tics of the product In fishery industry, many attempts have

been made to apply descriptive analysis in quality programs

of catfish products Yet, no standard application was done

Suvanich and Marshall (1998), instead of a fully trained

panel, used an experience panel with only 4 h of training to

examine the quality of catfish frames The authors also

adapted the terminology from studies of Huss (1988) and of

Pigott (1988), which focused more on detecting products’

spoilage Furthermore, fishery processors tended to use

descriptive analysis and expert tasters to examine sensory

acceptance/rejection as well as freshness/deterioration of

their products (Harewood 1996; Bal’a et al 1996; Mai et al.

2007) This approach, however, was inappropriate, as people

like food flavors, not discrete sensory properties (Prescott and

Murphy 2009) This inappropriateness has been long pointed

out in literature, resulting in a wide application of acceptance

testing with consumers (Sidel et al 1981; Lawless and

Heymann 2010; Prescott et al 2011).

One of the most common applications of descriptive

analysis is to develop flavor lexicon for food products, due to

the great benefit such lexicon can bring to the product

devel-opment and quality control Drake and Civille (2002) stated

that flavor lexicons can be used to record and define product

flavor, compare products, and determine storage stability, as

well as interface with consumer liking and acceptability and

chemical flavor data These benefits of sensory lexicon were

concluded in Lee and Chambers (2007) study of green tea A

number of lexicons have been recently developed for a variety

of food and nonfood products, Western European cheeses

(Talavera-Bianchi and Chambers 2008), lip products (Dooley

et al 2009), almond (Civille et al 2010), brewed coffee

(Hayakawa et al 2010a), French bread in Japan (Hayakawa

et al 2010b), fresh leafy vegetables (Talavera-Bianchi et al.

2010), beef (Adhikari et al 2011) to name but a few There-fore, a lexicon of pangasius sensory properties will definitely

help processors, researchers and consumers to identify the

common characteristics of pangasius products, acknowledge

the on/off-flavors and identify the properties that are appeal-ing to a specific market segment

Another issue is that consumers often have to make pur-chase decision for raw catfish fillets (included fresh, frozen or thawed) but consume the cooked ones Some attempts have been made to investigate the possibility of making prediction for quality of cooked product from quality of raw product

Rødbotten et al (2009) have productively applied descriptive

analysis to demonstrate the quality of raw salmon fillet as a predictor of cooked salmon quality Prediction models for cooked sensory attributes of shrimps from raw sensory

attributes were also made by Erickson et al (2007) These

accomplishments encourage a similar prediction could be made for catfish fillet products, to assist purchasers in making purchase decision

Given that, the present study aimed to develop a lexicon of

sensory properties for catfish (Pangasius hypophthalmus)

fillets, using QDA procedure This study also aimed to examine sensory characteristics of raw catfish fillets (fresh and frozen) as predictors of sensory properties of cooked

catfish fillets The findings were expected to allow pangasius processors and purchasers to select pangasius fillet sources

based on their products’ consistent sensory attributes and support the development of control plans before exportation,

or create innovative products with specific organoleptic

char-acteristics (Phan et al 2011) The lexicon, in addition, would

be greatly helpful to product developers, researchers and technologists in understanding the sensory characteristics of

pangasius fillets produced in Mekong Delta, the most

impor-tant region for catfish farming in Vietnam (Nguyen et al.

2007)

MATERIALS AND METHODS Samples

Samples used in this study were supplied by four different manufacturers in Mekong Delta and one manufacturer in Ho-Chi-Minh City in Vietnam These five factories were coded as Mekong1, Mekong2, Mekong3, Mekong4 and HCMC The samples included six raw fresh and 10 frozen products with different storage lengths of 3 months, 6 months and no storage Raw fresh fillets were evaluated in two sample types: raw and cooked (200C, 25 min, wrapped in aluminum foil) Frozen fillets were assessed in four sample types: frozen (evaluated right after taken from the freezer), thawing (45 min after taken from the freezer), thawed (completely

thawed at room temperature) and cooked (200C, 25 min,

wrapped in aluminum foil) Information about each fillet

product is shown in Table 1

Sample Preparation and Serving

(Codex Alimentarius 2001)

Raw Fresh Samples Five fresh samples, preserved in ice,

were delivered to the sensory laboratory in 10 h after filleting

at the suppliers’ production plants These samples were then

stored in a refrigerator (Toshiba GR-S22VPT/VT) at 0–4C

overnight On the test day, these fillets were taken out of the

fridge about 1 h before the evaluation to let the sample

tem-perature reach to the room temtem-perature, which was 22–25C

The sixth sample was purchased at a local market and was

filleted at the sensory laboratory on the day of testing This

sample was then served directly to the panel

Frozen Samples

These samples were received 24 h before evaluation and were

stored in a freezer (Sanyo Freezer HF-S6L) at (-18) C

over-night Frozen samples were evaluated immediately after taken

from the freezer The panel had to evaluate all six the

attributes of these samples within 5 min

Thawing Samples

The frozen samples were let to thaw at room temperature in

45 min At the point of evaluation, the sample’s ice layers on

both surfaces have melted completely, facilitating the obser-vation of color and fiber of the fillet

Thawed Samples

The frozen samples completely thawed at room temperature This process took about 1.5 h

Raw fresh, frozen, thawing and thawed samples were

with the appropriate three-digit code

Cooked Samples

This type of sample was prepared by cutting each fillet into three parts: head, middle and tail The middle part (about

50 mm length for big-size fillet) was wrapped in aluminum foil and cooked in an oven (Sanaky VH 50N, 1300 W) at 200C

in 25 min Cooked samples were served hot in disposable, odor-free, round plastic plates (d= 160 mm, THH Co LTD, Vietnam), coded by the random three-digit number

Descriptive Panel (ISO 8586-1, 1993)

A screening questionnaire has been distributed to the student, and faculties of two colleges The questionnaire asked about their food habit, any allergy they might have with fish and seafood products, if they smoke, health status and their avail-ability for the project (they need to have commitment for at least 6 months) Thirty-seven candidates (12 faculties and 25 students) were qualified to go to the next selection process

TABLE 1 DETAILED DESCRIPTION OF CATFISH

FILLET SAMPLES USED IN THE STUDY Sample name Type Sampling place

Storage Condition Storage time F-Mekong1 Fresh Cao Lanh – Mekong delta 0–4C 24 h

F-Mekong2 Fresh Sa Dec – Mekong delta 0–4C 24 h

F-Mekong3 Fresh Tien Giang – Mekong delta 0–4C 24 h

F-HCMC1 Fresh Ho Chi Minh city 0–4C 24 h

F-HCMC2 Fresh Ho Chi Minh city 0–4C 24 h

F-HCMC3 Fresh Ho Chi Minh city No storage a

Fr-Mekong1-0T Frozen Cao Lanh – Mekong delta -18C No storage b

Fr-Mekong1-3T Frozen Cao Lanh – Mekong delta -18C 3 months

Fr-Mekong1-6T Frozen Cao Lanh – Mekong delta -18C 6 months

Fr-Mekong2-0T Frozen Sa Dec – Mekong delta -18C No storage b

Fr-Mekong2-3T Frozen Sa Dec – Mekong delta -18C 3 months

Fr-Mekong2-6T Frozen Sa Dec – Mekong delta -18C 6 months

Fr-Mekong3-0T Frozen Tien Giang – Mekong delta -18C No storage b

Fr-HCMC1-3T Frozen Ho Chi Minh city -18C 3 months

Fr-Mekong4-3T Frozen Can Tho – Mekong delta -18C 3 months

Fr-Mekong4-6T Frozen Can Tho – Mekong delta -18C 6 months

a Sample was evaluated right after filleting, at ambient temperature.

b Due to transportation to the sensory laboratory, samples were, in fact, stored at (-18)C in about

12 h before evaluation.

This process included two test sessions, each comprised of

four tests The first session included basic taste detection test,

ranking test on sweetness, two ranking tests on color and odor

description The second session included ranking test on

umami taste, texture description and taste matching All

can-didates were ranked from first to 37th based on their total

scores in all eight tests Eleven panelists were chosen using

fol-lowing criteria: high total scores, not score zero in any test and

high score in odor and texture description tests

The panel had three 2 h sessions for product orientation,

each session dedicated to one of three types of samples: fresh,

frozen (included three states: frozen, thawing and thawed)

and cooked During the orientation, the panel manually

gen-erated as many as possible descriptors of odor, appearance

and texture for raw fresh and frozen samples For cooked

sample, the panelists opened the aluminum foil and sniffed

the sample to develop odor/aroma descriptors Then they

tasted the sample to develop descriptors for oral textural

properties, taste and mouthfeel The product frame of

refer-ence included neither catfish beyond the samples in the study

nor aged samples

After completing the orientation sessions, the panel

par-ticipated in a 55 h training course (about 8 h per week), to

improve their ability of detection and discrimination of

sensory properties of fresh, frozen and cooked fillets, based on

the descriptors generated above The panelists were trained to

estimate the intensity of the sensory attributes using a line

scale of 100 mm, anchored with “very weak” and “very

strong” at two ends The panelists’ performances were

moni-tored during the training phase and the testing phase The

panel was required to reach the consensus in understanding

and evaluating the attributes Panel agreement was evaluated

by examining the interaction between panelists and the

prod-ucts in an ANOVA model Panel’s repeatability was also tested

to assure that panelists could repeat their evaluation in

differ-ent testing sessions

Test Procedure

A descriptive analysis procedure was adapted as the test

method Evaluation took place at the Sensory Laboratory at

Ho Chi Minh City University of Technology,Vietnam Testing

environment was put under control for distracting factors

such as temperature (22–25C), lighting (fluorescent only),

odor (free of extraneous odor) and noise A total of six raw

fresh, 10 raw frozen and 16 cooked samples (made from both

raw and frozen fillets) were evaluated in three replicates,

resulting in 96 sample trials The panel, on average, completed

12 sample trials per one 2 h session In total, eight evaluation

sessions have been conducted In each evaluation session, the

panel was randomly served one sample at a time to estimate

the intensity of all sensory attributes of the sample on a line

scale of 100 mm, anchored by “very weak” and “very strong.”

References were also provided during testing The panelists tasted only the cooked sample to evaluate taste, mouthfeel and oral textural characteristics

Data Analysis

Multiple Factor Analysis (MFA) (Pagès 2004) was used to investigate the relationship between groups of properties of raw, frozen and cooked fillets Principal Component Analysis (PCA) was performed to the mean scores of sensory attributes of cooked samples to investigate the correlation between color attributes and flavor of cooked fillets These analyses were performed using R 2.10.0 (R Development Core Team 2007) and the SensoMineR (Lê and Husson 2008; Husson and Lê 2009) and FactoMineR packages (Husson

et al 2007; Lê et al 2008).

RESULTS AND DISCUSSIONS Sensory Lexicon of Catfish Fillets Produced in Mekong Delta of Vietnam

Three sensory lexicons have been developed to describe catfish fillets corresponding to a sequential process from raw

to frozen, thawing, thawed and cooked The descriptive pro-cedures, basically, followed those for the quantitative descrip-tive method (Stone 1992) All descriptors were relevant to products, redundancy eliminated and discriminative Defini-tion, references and evaluating procedure were also devel-oped for each attribute by agreement The validation of these lexicons has been confirmed by referring to precedent studies and Codex guidelines for the sensory evaluation of fish and shellfish in laboratories (Codex 2001) However, while Codex only gives example of general attributes for a broad fishery product category, our study has gone further and more spe-cific in describing all possible sensory aspects of catfish fillets, covered all phases of raw, frozen, thawing, thawed and cooked

Raw Fresh Samples

Raw fresh fillets were described by 22 descriptors, included five appearance properties, five colors, seven odors and five textural attributes (Table 2) The appearance attributes such

as smoothness and fat leftover reflected the professional of the filleting process Fresh catfish fillet was mainly white and pink, but some samples exposed yellow and orange The brown-red color was only distributed at the flesh along the backbone of the fish This was also an indicator for the fillet-ing professional since market fillet product should not contain any red meat Raw fresh catfish possessed aromatic notes associated with the growing (i.e., algae, sulfurous and ammonia) and processing (i.e., oxidized fat) conditions Raw

catfish fillet was identified as soft, elastic, cohesive and moist

in texture It was also characterized as firm This term,

however, was not opposite to soft Instead, these two terms

characterized two different textural dimensions Soft referred

to the ease of deformation, and firm related to force

resistance

Frozen Samples

Frozen ones were characterized by seven descriptors for frozen phase (mostly appearance and color) (Table 3), six descriptors for thawing phase (only appearance and color) (Table 4) and nine descriptors for thawed phase (only texture

TABLE 2 SENSORY ATTRIBUTES (n= 22) OF RAW FRESH CATFISH FILLETS THE SAMPLES WERE FILLETED AND KEPT IN ICE NO LONGER THAN 24 H BEFORE EVALUATION

Category Attribute Description of attribute Reference

Appearance Thickness Thickness of the fillet,

especially the middle part

NA Smoothness Smoothness of the surface Soft tofu

Fat-leftover Fat quantity along the spine

and edges of the fillet

Picture of fillet extreme in this attribute, in agreement of the panel

Fiber-curvature Curvature of fiber on both

faces of the fillet

Picture of fillet extreme in this attribute, in agreement of the panel

Fiber-sharpness Sharpness of fiber on both

faces of the fillet

Picture of fillet extreme in this attribute, in agreement of the panel

Color White The whiteness of the flesh White paper

Pink The pink color of the flesh Pink paint

Yellow The yellow color of the flesh Yellow paint

Reddish brown The brownness with some red Color of beef

Orange The orange color of the meat Orange juice

Odor Fishy Odor associated with aged

fish, as demonstrated by

trimethylamine

Fish liver oil

Algae Odor associated with the

blue-green algae growth in the pond water

Algae in pond

Raw clam Odor associated with smell of

raw clam

Raw clam Sulfurous Odor associated with sulfur

smell in boiled aged egg

10-day aged egg Ammonia Odor associated with

ammoniac gas

Ammonia Fat Odor associated with fish fat Fat taken from catfish belly

Oxidized fat Odor associated with oxidized

fish fat

Fish belly fat kept for about

5 h at room temperature Texture Soft Easy to be deformed Soft tofu

Elastic Mechanical textural attribute

relating to the rapidity of recovery from a deforming force

Hard tofu

Firm Textural attribute relating to

the force required to achieve a given penetration

of the fillet

Jelly candy

Cohesive Textural attribute relating to

the degree to which the fillet can be deformed before it breaks

Hard tofu

Moist Water content in the fillet Soft tofu

NA, Non applicable.

was described) (Table 5) At frozen phase, the fillets did not

expose any aromatic notes but fishy Since fillet was covered by

ice, colors were barely exposed At thawing phase, the ice

surface of the fillet has melted completely, facilitating color

and fiber evaluation However, the fillet was still very firm and

cold, which suppressed any noticeable odor Color and

appearance attributes of thawing phase were also found in

thawed phase Frozen catfish fillet, after thawed, possessed

similar aromatics and textural properties to raw fresh fillets

However, aged-frozen fillets were perceived stronger in odors

such as fishy, ammonia, sulfur and softer and less elastic than

raw fresh fillets

Cooked Samples

Cooked samples were profiled by 18 descriptors, included

four colors, five odors, five oral textural properties and three

tastes/mouthfeel (Table 6) Some of the attributes were found

in both odor/aroma and flavor in the mouth, i.e., fishy, fat and

boiled chicken The panel has come to agree to list those

attributes in the odor/aroma category, since those attributes

were more intense when perceived orthonasally than

retrona-sally Boiled shellfish, boiled chicken, boiled egg, fat aromatic

and ammonia were perceived strongest when the aluminum

foil was just opened About 10 min later, only fishy remained

In general, cooked fillets were mild in flavor and aromatic

They were slightly salty but more “savory” (umami taste) and

fatty Boiled chicken was the most intense note in terms of intensity comparing to other aromatics Catfish were also found to be soft, moist, flaky and slightly cohesive but less chewy in texture Chambers and Robel (1993) have developed

an extensive flavor and texture lexicon for freshwater fish including pond-raised and raceway-raised channel catfish from Kansas and Mississippi The authors reported that catfish has a full base flavor in which earthy is one of the key-notes This attribute, however, was not present in our lexicon for cooked fillet Instead, boiled-egg note was reported This note associated with the sulfurous smell in boiled-aged egg It was reminiscent of decaying vegetation note in the “muddy” flavor group reported in Chambers and Robel (1993) Other similarity was also found, such as boiled-chicken note in our lexicon could be related to corn-like, nutty/buttery, white meat and sweet aromatic; fat aromatic related to oily aro-matic; fishy to fresh fish in Chambers and Robel’s lexicon Besides, similar textural terms, such as flaky, juicy, cohesive, chewy (chew count) were reported in both studies Our lexicon was also consistent with the lexicon of pond-raised

catfish documented in Johnsen et al (1987), which

com-prised similar descriptors such as boiled chicken, cardboard (oxidized fats and oils),and salty The fish in our study pos-sessed sensory properties similar to southern flounder,

reported in Drake et al (2006) Both catfish and flounder

pos-sessed aromatics associated with chicken broth, fresh fish, earthy, fat and slightly salty The only common in odor

TABLE 3 SENSORY ATTRIBUTES (n= 7) OF

FROZEN CATFISH FILLETS Category Attribute Definition Reference

Appearance Equality of glaze The equality of ice on both

sides of the fillet

Picture of fillet extreme in this attribute, in agreement of the panel

Quantity of ice Quantity of ice flakes on both

sides of the fillet

Picture of fillet extreme in this attribute, in agreement of the panel

Fiber-sharpness Sharpness of the fiber on both

sides of the fillet

Picture of fillet extreme in this attribute, in agreement of the panel

Odor Fishy Odor associated with aged

fish, as demonstrated by

trimethylamine

Fish liver oil

TABLE 4 SENSORY ATTRIBUTES (n= 6) OF

THAWING CATFISH FILLETS Category Attribute Definition Reference

Appearance Fiber density Density of fiber on the fillet Catfish fillet

Fiber curvature Curvature of fiber Catfish fillet Color Red Redness of the flesh along the spine Red

TABLE 5 SENSORY ATTRIBUTES (n= 10)

OF THAWED CATFISH FILLETS (THAWED AT ROOM TEMPERATURE)

Category Attribute Definition Reference

Odor Fishy Odor associated with aged

fish, as demonstrated by

trimethylamine

Fish liver oil

Sulfurous Odor associated with sulfur

smell in boiled-aged egg

Boiled-aged egg Ammonia Odor associated with

ammonia

Ammonia Fat Odor associated with fish fat Fat in the belly of catfish

Texture Soft Easy to be deformed Soft tofu

Elastic Mechanical textural attribute

relating to the rapidity of recovery from a deforming force

Hard tofu

Firm Textural attribute relating to

the force required to achieve a given penetration

of the fillet

Jelly candy

Cohesive Textural attribute relating to

the degree to which the fillet can be deformed before it’s broken

Boiled chicken meat

Moist Water-content in the fillet Soft tofu

Smooth Smoothness of the fillet

surfaces

Soft tofu

TABLE 6 SENSORY ATTRIBUTES (n= 18) OF COOKED CATFISH FILLETS SAMPLES WERE WRAPPED IN ALUMINUM FOIL AND COOKED IN OVEN AT 200C IN 25 MIN

Odor/aroma Boiled shellfish Odor associated with boiled shellfish Boiled shellfish

Boiled chicken Odor associated with boiled chicken Chicken stock Boiled egg Odor associated with boiled-aged egg Boiled-aged egg Fat Odor associated with cooked fish fat, evoke greasy perception Cooked fish fat Fishy Odor associated with aged fish, as demonstrated by

trimethylamine

Fish liver oil Ammonia Odor associated with ammonia Chinese steamed bun (baozi) Oral textural properties Soft Easy to be deformed Soft tofu

Flaky Capacity of easily forming flake of the fiber NA Chewy Numbers of times required to chew a piece of sample to the

state ready to swallow

Boiled chicken, 2 ¥ 2 cm cube Cohesive Textural attribute relating to the degree to which the fillet can be

deformed before it breaks

Boiled chicken

Taste Umami The taste on the tongue associated with glutamate ions Monosodium glutamate (MSG)

Salty The taste on the tongue associated with sodium ions Monosodium chloride Mouthfeel Fatty The mouthfeel described as greasy and associated with fat Boiled catfish fat

NA, non applicable.

between catfish in our study and farmed cod reported in

Sveinsdóttir et al (2010) was the aroma associated with

trim-ethylamine and sulfur However, these two fish were similar in

many of textural characteristics In addition, sensory

proper-ties of catfish were fundamentally different from those of

salmon (Green-Petersen et al 2006) and ocean fish due to its

white meat character (Chambers and Robel 1993)

In aquaculture quality control, it’s very important to detect

off-flavors of a fishery product, since this defect jeopardizes

consumer satisfaction and future purchase Therefore,

projects in aquaculture often focus on testing and managing

off-flavors in fishery products Van der Ploeg (1991), from the

perspective of flavor quality control, has developed a flavor

wheel for channel catfish, consisting of six categories of flavor

descriptors, i.e., acceptable (such as corn, nut-like, buttery/

fat, chicken), blue green algae (woody, musty, etc.), chemical

(metallic, diesel, etc.), decay (egg-sulfuric, rotten, etc.),

veg-etable (grassy, mushroom, etc.) and fishy (fish oil, rancid,

stale, etc.) The author reported that blue green algae,

chemi-cal and decay were three objectionable categories for channel

catfish Depending on intensity, flavor descriptors in fishy and

vegetable categories could be acceptable or objectionable to

the consumers Another study of Tucker and van der Ploeg

(1999) on pond-raised catfish also reported common

off-flavors in catfish consisting of blue-green algae, decay/rotten

and petroleum Adopt the classification of van de Ploeg, in

this study, we considered algae, boiled egg, sulfurous, oxidized

lipid, ammonia and fishy as objectionable flavors Boiled

chicken, boiled shellfish, boiled fat, clam and lipid smell were

on-flavors of the catfish fillets Other attributes, depending on

their intensity, were acceptable or objectionable Due to its

validity and completion of describing catfish fillets, this

lexicon could be a reference for Vietnamese fish producers

and purchasers to establish a quality control program for

catfish fillets

Common Structure between Catfish Fillets in

Two Phases: Raw Fresh and Cooked

As one of our main objectives was to identify raw fresh fillets

sensory descriptors that could be used to predict cooked fillets

sensory descriptors, we performed an MFA on both types of

descriptors The first MFA maps of individuals and variables

for fresh samples are shown in Figs 1 and 3 The first factor

was highly correlated with the variables belong to the two

groups This axis opposed samples F-HCM2, F-HCMC3 to

F-Mekong1, F-Mekong2, F-HCMC1 and F-Mekong3 This

first MFA factor opposes also the descriptors boiled fat, greasy

mouthfeel, umami, elastic, cohesive to the descriptors

boiled-aged egg, oxidized lipid, ammonia, sulfur In the literature, the

higher (resp lower) the value of the descriptors of the first

(resp second) group, the better (resp the poorer) the quality

of the catfish product This first axis can be interpreted as a

gradient of quality The two products F-HCMC2 and F-HCMC3 were firm, cohesive, chewy and flavored-boiled fat and chicken; whereas the others possessed the smell of oxi-dized lipid, sulfur and ammonia in raw phase; soft and smelled of boiled-aged egg in cooked phase

It was noticed that colors such as pink, yellow and orange were found to have correlation with off-flavors (Fig 2) Pink highly correlated with oxidized lipid, sulfurous and ammonia; while yellow and orange well correlated with algae smell However, white did not show any relationship with the on-flavors of both raw-fresh and cooked samples This result

was consistent with the finding of Rødbotten et al (2009) that

color had less impact on the quality of cooked salmon sample These findings casted doubt on the rule of thumb of using color to judge the quality of catfish fillet, which has been very common in catfish quality assurance and trading

Factors from separate PCA of two groups are presented in Fig 3 The figure showed a strong correlation between the first factor of MFA and the first factor of each separate analy-sis These factors of separate analyses were also highly corre-lated to one another The same observation can be made for the second factor This shows that the first map of MFA was similar to the one of each separate analysis In other words, this suggested a possibility to obtain one map from the other

by slight rotation This finding supported our assumption of

a feasible prediction of sensory properties of catfish fillets at cooked phase from fresh phase and vice versa Representation

of groups was performed to give a better demonstration of the common structure between the two groups of variables (Fig 4) The two groups of raw fresh and cooked were almost identical on the map They were strongly related to MFA factor 1, which meant that many of variables of these two groups were related to this factor This result, once again,

FIG 1 FIRST FACTORIAL MAP FROM MFA FOR FRESH SAMPLES

Dim 2 (21

Dim 1 (47.32%) –1.0

–0.5

RawFresh (f)

moist_f ammonia_f smoothness_f

ammonia_c

thickness_f

algae_f

firm_f

elastic_fgrey_ccohesive_f fishy_c yellow_c yellow_f orange_f

boiled fat_c cohesive_c chewy_c fatty.mouthfeel_c clam_f

boiled.chicken_c

white_c white_f

curve.of.fiber_f

fat.leftover_f

umami_c reddish.brown_f

salty_c flaky_c

lipid_f moist_c fishy_f

soft_c sulfur_f pink_f

pink_c boiled.egg_c

boiled.shellfish_c

soft_f oxidized.lipid_f

cleamess.of.fiber_f

CookedFresh (c)

0.0 0.5 1.0 FIG 2 FIRST FACTORIAL MAP FROM MFA FOR

FRESH SAMPLES: VARIABLES AT TWO PHASES: RAW FRESH (f) AND COOKED (c)

Dim 1 (47.32%) –1.0

Dim1.CookedFresh

Dim4.CookedFresh Dim5.CookedFresh Dim3.CookedFresh

Dim2.RawFresh

Dim3.RawFresh

Dim5.RawFresh

Dim4.RawFresh

Dim1.RawFresh CookedFresh

FIG 3 REPRESENTATION OF THE FACTORS

FROM SEPARATE ANALYSES (PCA) OF TWO GROUPS: FRESH AND COOKED

stabilized the possibility of making prediction of sensory

properties of cooked fillet from sensory properties of raw

fresh fillets

Common Structure between Sensory

Characteristics of Catfish Fillets in Four

Phases: Frozen, Thawing, Thawed and

Cooked

The relationship among the sensory properties of catfish

fillets at four different phases, i.e., frozen, thawing, thawed

and cooked was another main focus of this study In other

words, the feasibility of making mutual prediction among the

four was investigated An MFA was also performed on those

four types of properties Figures 5 and 6 showed the

represen-tation of individuals (products) and variables (attributes)

from MFA for frozen samples The first axis was highly

corre-lated with the variables of two groups of thawed and cooked

Those variables were attributes of odors and textures

Accord-ing to these two figures, the catfish fillets coded Mekong1, and

Mekong3 were “firm,” “elastic” and “cohesive” at the thawed

state; have smells of “boiled chicken,” “boiled shellfish,”

“boiled fat” as well as cohesive texture and umami taste when

cooked These properties were considered on-flavors Fillets

from Mekong2 and Mekong4 were “smooth,” “soft” and

“ammoniac” at thawed phase; possessed “boiled aged egg,”

“ammonia” smells and “salty” once cooked These were

regarded as off-flavors This axis can also be interpreted as a

gradient of quality The second axis was contributed by the

variables of frozen and thawing groups, with the dominant

contribution of color attributes This axis opposed the sample

“Mekong1-6T” to the rest This product Mekong1-6T was

perceived to have bright colors of “orange,” “pink,” while the others were “white” and “yellow.”

Length of storage, statistically, did not have any influence

on the sensory quality of the samples, especially at thawed and cooked phases However, from observation, the appearances

of the fillets with 6 months of storage were not as good as the ones with shorter length of storage Furthermore, there was

no clear evidence of origin-resulting differences among the samples

Factors from separate analyses, which were performed on four types of catfish descriptors, were presented in Fig 7 As observed from the plot, the first factor of MFA was highly cor-related with the first factors of separate analysis of thawed, cooked and frozen groups These factors of separate analyses were highly correlated with each other The second MFA factor was only correlated with the second factor of frozen group However, there were strong correlations between the first factor of MFA with the second factor of thawing group; and between the second factor of MFA with the first factor of thawing group This suggested that the first MFA map was roughly similar to the one of frozen group with a slight rota-tion The first MFA map could also be derived from the PCA map of thawing group by counterclockwise rotation and flip First factor was the only common between MFA map and the separate maps of thawed and cooked groups This factor, however, was the only factor significantly distinguishing the individuals (products) Thus, it could be said that the MFA map was also similar to separate ones of thawed and cooked in the sense of discrimination of the products

Valid predictions can be made for cooked from thawed due

to the strong relationship one another (Fig 6) Thawed fillets, which were elastic, firm and cohesive, tended to possess on-flavors such as chicken, shellfish, fat smells and umami

FIG 4 GROUPS REPRESENTATION FOR FRESH CATFISH FILLET SAMPLES

FIG 5 FIRST FACTORIAL MAP FROM MFA: MEAN INDIVIDUALS

(FROZEN PRODUCTS)