MULTIVARIATE QUALITY PREDICTION OF COD (GADUS MORHUA) AND HADDOCK FILLETS (MELANOGRAMMUS

AEGLEFINUS) STORED UNDER SUPERCHILLING AND TEMPERATURE ABUSIVE CONDITIONS

Gudrun Olafsdottir*, Hélène L. Lauzon, Rósa Jónsdóttir, EmilíaMartinsdóttir,

Icelandic Fisheries Laboratories, Skulagata 4, 101 Reykjavik, Iceland

2nd Workshop "Cold-Chain-Management”. Bonn,8-9th May 2006

OBJECTIVES

to achieve better understanding better understanding of the effect of temperature effect of temperature on the complex spoilagecomplex spoilage changes of fish

volatile compounds as quality indicatorsquality indicatorsGCGC (gas chromatography) analysischaracterize the spoilage potential of specific spoilage organisms (SSO) specific spoilage organisms (SSO)

to apply an electronic noseelectronic nose as a rapid rapid technique to monitor classes of volatile compounds as indicators of quality changesquality changesduring storage of chilled fish

storage studies of chilled cod and haddock fillets – natural productsexplore the correlation of microbial, chemical, sensory and emicrobial, chemical, sensory and e--nose nose data by using multivariate models (PLSR / PCA (PLSR / PCA -- SIMCA)SIMCA)predict the quality or classify products according to sensory criteriasensory criteria.

Various factors influence the spoilage of fish

Species cod and haddock

Seasonal condition autumn

Fishing grounds Northeast and Southwest Iceland

Catching methods longline / bottom trawl

HandlingFlake ice / slurry iceSuperchilling CBC (contact blast and cooling)

Processing skinless fillets

Storage techniques styrofoam boxes

Time / temperature -1,5 to 15°C

Protein / i.e. sarcoplasmic proteins → peptides→ amino acids

Lipids / i.e. phospholipids → PUFA

Soluble substances in the muscleNucelotidesNPN non protein nitrogenous components TMAO → TMA/DMA, FA, pH ↑

Glycolysis pH↓, Lactic acid↑

Rigor mortis ATP→ Inosine →Hx →Urea

Autolysis

Endogenous enzymes lipoxygenase, cathepsin, calpain

Microbial growthSSOSpecific SpoilageOrganisms

Oxidationpro- and antioxidants

Fresh odor

Spoilage odor

Complex spoilage changes in fish

Texture firm – soft

Color changes

Identify Identify qualityquality indicatorsindicators

QUALITY INDICATORS in chilled fishSpoilage odors Volatile microbial metabolites Specific Spoilage Organisms (SSO)

Ammonia odor, dried fish, old fish

odorAmines: TMA/DMA

NH3

Putrid odor, onion like, old cabbage,

rotten egg

Sulfur compounds

SSOs Pseuodomonas spp. P. phosphoreum H2S producers

Sweet - sour, malty and fruity odors

Alcohols, aldehydes, ketones, esters

METHODS – analysis of volatile compounds

Electronic nose FreshSenseMaritech, Iceland (prototype)Electrochemical sensors

CO, NH3, H2S, SO2

2.6L sampling vesselPump Continuous sampling for 5 minutesApprox. 500 g fillets Temperature monitoring /control

Rapid techniqueA few sensors sensitive to the main classes of compounds produced in chilled fish

Comparison with gas chromatographyIdentify the main classes of compounds produced during storage to select appropriate sensorsGC-MS / GC-O

Storage study of haddock filletsComparison of the e-nose sensors and GC analysis of the main classes of volatile compounds produced

during chilled storage

0 5 10 15Days

arbi

trary

uni

ts

Alcohols, aldehydes and esters

Amines (TMA)

Sulfur compounds

0

400

800

1200

0 5 10 15Days

Res

pons

e se

nsor

s (n

A)

CO

SO2

NH3

CO sensor CO sensor alcoholsalcohols, , aldehydesaldehydes and esters and esters

SOSO22 sensorsensor sulfursulfur compoundscompounds

NHNH3 3 sensor sensor aminesamines

Electronic nose sensors Gas chromatography

In: Olafsdottir G. 2003. Developing rapid olfaction arrays for determining fish quality. In Ibtisam E Tothill(Ed) RAPID AND ON-LINE INSTRUMENTATION FOR FOOD QUALITY ASSURANCE. WoodheadPublishing Ltd, Cambridge, England, pp. 339-360

Examples of the dynamic evolution of volatile compounds in spoilage of fish (cod fillets 0°C)

0

5

10

15

20

25

30

35

40

45

50

0 2 4 6 8 10 12 14 16

Days from catch

PAR

0

100

200

300

400

500

600

700

800

900

1000

PAR

(iso

buta

nol)

ethanol

3-methyl-1-butanol

2,3-butandiol

ethyl acetate

isobutanol

End of shelf-life

0

5

10

15

20

25

30

35

40

45

50

0 2 4 6 8 10 12 14 16

Days from catch

PAR

0

50

100

150

200

250

300

PAR

(ace

toin

)

6-me-5-hepten-2-one

2-butanone

3-pentanone

acetic acid

acetoin

End of shelf-life

Alcohols, esters Ketones, acids

From: Olafsdottir, G., Jonsdottir, R., Lauzon, H.L., Luten, J., Kristbergsson, K. 2005. Characterization of volatile compounds in chilled cod (Gadus morhua) fillets by gas chromatography and detection of quality indicators by an electronic nose. JAFC, 53 (26), 10140 -10147.

Potential quality indicators in cod fillets (0°C)

0

50

100

150

200

250

300

350

400

450

0 2 4 6 8 10 12 14 16

Days from catch

PAR

- TV

B-N

- se

nsor

s

6,2

6,3

6,4

6,5

6,6

6,7

6,8

6,9

7,0

pH

TMATVB-NCO sensorNH3 sensoracetoinpH

End of shelf-life

1

2

3

4

5

6

7

8

9

0 2 4 6 8 10 12 14 16

Days from catch

log

CFU

/g

TVC

H2S counts

Pseud

Pp

End of shelf-life

Microbial counts: TVC (total viable counts)H2S producersPseudomonas spp.P. phosphoreum (predominating)

Volatiles: TMA (trimethylamine), TVB-Nacetoin (3-hydroxy-2-butanone)

E-nose: CO and NH3 sensor

Quantification and identification of the main classes of compounds in cod fillets during storage (0°C)

by GC-MS and GC-O

0

10

20

30

40

50

60

70

80

90

100

110

120

Ketones TMA Alcohols Acids Aldehydes Esters Sulfurcompounds

PAR

Day 4 Day 7 Day 10 Day 12

3-methyl-butanal (sweet, malty odor)hexanal (grass),heptanal (boiled potato),nonanal, decanal (fresh, floral, sweet)

acetoin, 3-pentanone, (sweet, butter-like, sour)

Fishy odorDMS, DMDS, DMTS: onion like

rotten, sulfur, cabbage

Sickenly sweet odor

33%

29%

4% 3%

15%

<1%

ethanol, 3-methyl-1-butanol, 2-methyl propanol

No odor detected

Early spoilage Early spoilage changeschangesKetonesKetonesAlcoholsAlcoholsAldehydesAldehydes

Late spoilage changesLate spoilage changesTMA TMA EstersEstersAcidsAcids(Sulfur(Sulfur compounds)compounds)

Sensory rejection on day 12

Storage studies - METHODSFive storage studies on cod and haddock fillets packed in styrofoam boxes

from two factories in Icelanddifferent temperatures (-1,5 to 15°C)

Traditional reference methodsSensory analysis

Torry scoreMicrobial counts

TVC (total viable counts) (15°C)SSO (specific spoilage organisms)

– Pseudomonas spp– H2S-producers– Photobacterium phosphoreum

Chemical analysis TVB-N (total volatile basic nitrogen)

Electronic nose

Data analysis

Multivariate modelsPLSR / PCA /SIMCA

Unscrambler Camo

Shelf-life determined by sensory analysis

3

4

5

6

7

8

9

10

0 2 4 6 8 10 12 14Days from catch

Torr

y sc

ore

0°C0°C a0°C+abuse7°C15°C

Sensory rejection Torry = 5,5

Influence of different storage temperature on the shelf-life of haddock fillets

From: Olafsdottir G, Lauzon HL, Martinsdottir E, Kristbergsson K. 2005. Influence of storage temperature on microbial spoilage characteristics of haddock fillets (Melanogrammus aeglefinus) evaluated by multivariate quality prediction. In press.

Storage conditions:

Study 1: 0°C, 7°C, 15°C

Study 2: 0°C, 0°C + abuse

2

3

4

5

6

7

8

0 2 4 6 8 10 12 14

days from catch

Pseu

dom

onas

spp

log

cfu/

g

15°C

7°C

0°C+ abuse0°C

0°C

0

100

200

300

400

500

600

700

800

0 2 4 6 8 10 12 14

days from catch

Res

pons

e C

O s

enso

e (n

A)

0°C

15°C 7°C

0°C

0°C+ abuse

Microbial counts Pseudomonas spp.

0

20

40

60

80

100

120

0 2 4 6 8 10 12 14

days from catch

TVB-

N m

g N

/100

g

7°C

15°C0°C+ abuse

0°C

0°C

E- nose: CO sensor

Influence of different storage temperature (0°C, 7°C, 15°C and 0°C, 0°C + abuse)

on the rate of spoilage changes haddock fillets

From: Olafsdottir G, Lauzon HL, Martinsdottir E, Kristbergsson K. 2005. Influence of storage temperature on microbial spoilage characteristics of haddock fillets (Melanogrammus aeglefinus) evaluated by multivariate quality prediction. In press.

TVB-N

Quality indictors do not always agree on the spoilage rate of sample groups

Partial Least Squares Regression (PLSR) modelPartial Least Squares Regression (PLSR) modelto explore the correlation of the variablesto explore the correlation of the variables

Selection of quality indicators [X] to predict sensory quality [Y] of haddock fillets (stored at different temperatures 0°C, 7°C, 15°C - 0°C, 0°C + abuse)

Quality indicators [X]H2S sensor - Tacc

NH3sensor – TVB-N - pH

CO sensor

Pseudomonas spp.H2S producersP. phosphoreumTVC

(Total viable microbial counts)

Different models were explored to select the best quality indicators Best models when all significant variables were used Of interest to use rapid techniques like the e-nose sensors (CO, H2S and NH3)

Sensory Quality [Y]Torry score

Partial Least Squares Regression (PLSR) model Partial Least Squares Regression (PLSR) model for haddock samples for haddock samples (0°C, 7°C, 15°C (0°C, 7°C, 15°C -- 0°C, 0°C + abuse0°C, 0°C + abuse)

Models based on pseudomonads counts and the CO, NH3 H2S sensors, and Tacc (time / temp) gave best results to predict sensory scores

(r2=0.94; RMSEP: 0,49)

5% error between predicted sensory scores and experimental values when using a subset of the data

Low temp. (0°C)

High temp. ( 7-15 °C)

PC1 TIME

PC2 TEMPERATURE

Storage studies on cod fillets at low temperature -1.5 (superchilling) to 0.5°C and temperature abuse

Factory 1: Process 1 Traditional (no cooling of fillets) + abuseProcess 2 Superchilling (CBC)

Factory 2: Process 3 Traditional (cooling of fillets)

Important to study natural products Catching, handling and storage conditions were not the sameThe aim was to select the best quality indicators (microbial, chemical, e-nose) to predict the sensory quality

-2-10123456789

10

D A B C E G F I H

Tem

pera

ture

°C

Initial temperature Average temperature of fillets

B 11.5dA 12 d

C 12.5 d

I 10.5 d

H 13.5 d

Process 3 (ice)

E 14.5 dG 15 d F 15+d

Process 2 (ice slurry)

Factory I Factory II

Traditional – no cooling

Superchilling - CBC

!16°C 8h

-1.5°C0.5 °C

0.5°C

!16°C 8h

!RT 16h

0.5 °C

!

17°C 16h

Process 1 (ice)

0.5 °C

CBC

D 8d

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Day

s fro

m c

atch

Temperature of fillets in experimental groups

Traditional - cooling

Influence of different processes and temperature conditions (-1.5 °C superchilling, 0.5 °C and abuse) on the shelf-life of cod fillets

Partial Least Squares Regression (PLSR) modelto explore the potential of the quality indicators (X)

to predict sensory quality (Y)

Cod fillets stored at -1.5 (superchilling) to 0.5°C + abuse– different factories and processes

Quality indicators [X]CO sensorTVB-N

P.phosphoreumTVCpseudomonadsH2S-producers

Sensory Quality [Y]Torry score

From: Olafsdottir G, Lauzon HL, Martinsdottir E, Oehlenschläger J, Kristbergsson K. 2006. Evaluation of shelf-life of superchilled cod (Gadus morhua) fillets and influence of temperature fluctuations on microbial and chemical quality indicators. J. Food Sci 71 (2): 97-109.

PLSR model for chilled and superchilled cod fillets

Multiple quality criteria based on 5 variables, the SSOs (P. phosporeum,H2S-producers, pseudomonads), CO sensor and TVB-N was needed when using a global model to classify samples from different factories stored at different temperatures

Process 3 – Cooling: H, I

Process 2 – Superchilling: E, F, G

Process 1 – Temp. fluctuations: A, B, D, C

• 85% correct classification based on sensory criteria (Torry score 7)

Volatile compoundsVolatile compounds contributed by different SSOsSSOs give information about the spoilage processes in fish

Electronic noseElectronic nose is a multimulti--indicator indicator device selective sensorsselective sensors for ketonesketones, amines, alcohols, , amines, alcohols, aldehydesaldehydes, acids, , acids, esters and sulfur compoundsesters and sulfur compoundse-nose can give more information than a single reference measurement for example TVB-N.

Establishment of quality Establishment of quality criteracritera or fixed values to determine the end of shelf-life or the quality of fish based on electronic nose responses, microbial counts and TVB-N reference methods will have to be developed for each product and the respective storageconditions.

CONCLUSIONS

FUTURE PERSPECTIVESImplementation of the electronic nose technique for quality

monitoring of fish products? The fast development of sensor technologies and data processing techniques will improve the possibility of quality monitoring in the food industry.

On-line monitoring ? Handheld devices ? “Smart” sensors in packaging?

samplingtemperature control exclusion of background odors and contaminantssensitive / selective sensors for quality indicating compoundsrapid detection of SSOs specific applications - establishment of quality criteria for different products and temperature conditions

AcknowledgementsFunding:

Icelandic Centre for ResearchAVS Research Fund of the Ministry of Fisheries

Participating companies:MaritechTrosTangiSkaginn (www.skaginn.is/)