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, Emilía Martinsdóttir, Icelandic Fisheries Laboratories, Skulagata 4, 101 Reykjavik, Iceland 2nd Workshop "Cold-Chain-Management. Bonn,8-9th May 2006
OBJECTIVES to achieve better understanding of the effect of temperature on the complex spoilage changes of fish volatile compounds as quality indicators GC (gas chromatography) analysis characterize the spoilage potential of specific spoilage organisms (SSO) to apply an electronic nose as a rapid technique to monitor classes of volatile compounds as indicators of quality changes during storage of chilled fish storage studies of chilled cod and haddock fillets natural products explore the correlation of microbial, chemical, sensory and e-nose e data by using multivariate models (PLSR / PCA - SIMCA) predict the quality or classify products according to sensory 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 Handling Flake ice / slurry ice Superchilling CBC (contact blast and cooling) Processing skinless fillets Storage techniques styrofoam boxes Time / temperature -1,5 to 15 C
Complex spoilage changes in fish Identify quality indicators Fresh odor Spoilage odor Endogenous enzymes lipoxygenase, cathepsin, calpain Glycolysis Rigor mortis Autolysis ph, Lactic acid ATP Inosine Hx Urea Oxidation pro- and antioxidants Microbial growth SSO Specific Spoilage Organisms Protein / i.e. sarcoplasmic proteins peptides amino acids Lipids / i.e. phospholipids PUFA Soluble substances in the muscle Nucelotides NPN non protein nitrogenous components TMAO TMA/DMA, FA, ph Texture firm soft Color changes
QUALITY INDICATORS in chilled fish Spoilage odors Volatile microbial metabolites Specific Spoilage Organisms (SSO) Sweet - sour, malty and fruity odors Alcohols, aldehydes, ketones, esters Ammonia odor, dried fish, old fish odor Amines: TMA/DMA NH 3 Putrid odor, onion like, old cabbage, rotten egg Sulfur compounds SSOs Pseuodomonas spp. P. phosphoreum H 2 S producers
METHODS analysis of volatile compounds Electronic nose FreshSense Maritech, Iceland (prototype) Electrochemical sensors CO, NH 3, H 2 S, SO 2 2.6L sampling vessel Pump Continuous sampling for 5 minutes Approx. 500 g fillets Temperature monitoring /control Rapid technique A few sensors sensitive to the main classes of compounds produced in chilled fish Comparison with gas chromatography Identify the main classes of compounds produced during storage to select appropriate sensors GC-MS / GC-O
Storage study of haddock fillets Comparison of the e-nose sensors and GC analysis of the main classes of volatile compounds produced during chilled storage Electronic nose sensors Gas chromatography 1200 CO SO2 Alcohols, aldehydes and esters Response sensors (na) 800 400 NH3 arbitrary units Amines (TMA) Sulfur compounds 0 0 5 Days 10 15 0 5 10 15 Days CO sensor alcohols, aldehydes and esters SO 2 sensor sulfur compounds NH 3 sensor amines 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. Woodhead Publishing Ltd, Cambridge, England, pp. 339-360
Examples of the dynamic evolution of volatile compounds in spoilage of fish (cod fillets 0 C) Alcohols, esters Ketones, acids PAR 50 45 40 35 30 25 20 15 10 5 0 ethanol 3-methyl-1-butanol 2,3-butandiol ethyl acetate isobutanol 1000 900 800 700 600 500 400 300 200 100 0 0 2 4 6 8 10 12 14 16 Days from catch End of shelf-life PAR (isobutanol) PAR 50 45 40 35 30 25 20 15 10 5 0 6-me-5-hepten-2-one 2-butanone 3-pentanone acetic acid 300 250 200 acetoin 150 100 50 0 0 2 4 6 8 10 12 14 16 Days from catch End of shelf-life PAR (acetoin) 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) Volatiles: TMA (trimethylamine), TVB-N acetoin (3-hydroxy-2-butanone) E-nose: CO and NH 3 sensor Microbial counts: TVC (total viable counts) H 2 S producers Pseudomonas spp. P. phosphoreum (predominating) PAR - TVB-N - sensors 450 400 350 300 250 200 150 100 TMA TVB-N CO sensor NH3 sensor acetoin ph 7,0 6,9 6,8 6,7 6,6 6,5 6,4 ph log CFU/g 9 8 7 6 5 4 3 TVC H2S counts Pseud Pp 50 6,3 2 0 6,2 0 2 4 6 8 10 12 14 16 Days from catch End of shelf-life 1 0 2 4 6 8 10 12 14 16 Days from catch End of shelf-life
Quantification and identification of the main classes of compounds in cod fillets during storage (0 C) by GC-MS and GC-O Sensory rejection on day 12 acetoin, 3-pentanone, (sweet, butter-like, sour) Early spoilage changes Ketones Alcohols Aldehydes PAR Late spoilage changes TMA Esters Acids (Sulfur compounds) 120 110 100 90 80 70 60 50 40 30 20 10 0 Day 4 Day 7 Day 10 Day 12 3-methyl-butanal (sweet, malty odor) hexanal (grass), heptanal (boiled potato), nonanal, decanal (fresh, floral, sweet) Ketones TMA Alcohols Acids Aldehydes Esters Sulfur compounds Fishy odor 33% 29% 15% Sickenly sweet odor ethanol, 3-methyl-1-butanol, 2-methyl propanol No odor detected 4% 3% <1% DMS, DMDS, DMTS: onion like rotten, sulfur, cabbage
Storage studies - METHODS Five storage studies on cod and haddock fillets packed in styrofoam boxes from two factories in Iceland different temperatures (-1,5 to 15 C) Traditional reference methods Sensory analysis Torry score Microbial counts TVC (total viable counts) (15 C) SSO (specific spoilage organisms) Pseudomonas spp H 2 S-producers Photobacterium phosphoreum Chemical analysis TVB-N (total volatile basic nitrogen) Electronic nose Data analysis Multivariate models PLSR / PCA /SIMCA Unscrambler Camo
Influence of different storage temperature on the shelf-life of haddock fillets Shelf-life determined by sensory analysis Storage conditions: Study 1: 0 C, 7 C, 15 C Study 2: 0 C, 0 C + abuse Torry score 10 9 8 7 6 0 C 0 C a 0 C+abuse 7 C 15 C 5 4 3 Sensory rejection Torry = 5,5 0 2 4 6 8 10 12 14 Days from catch 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.
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 Microbial counts Pseudomonas spp. E- nose: CO sensor TVB-N Pseudomonas s pp log cfu/g 8 7 6 5 4 3 15 C 0 C+ abuse 7 C 0 C 0 C Response CO sensoe (na) 800 700 600 500 400 300 200 100 15 C 0 C+ abuse 7 C 0 C 0 C TVB-N mg N/100g 120 100 80 60 40 20 15 C 7 C 0 C+ abuse 0 C 0 C 2 0 2 4 6 8 10 12 14 0 0 2 4 6 8 10 12 14 0 0 2 4 6 8 10 12 14 days from catch days from catch days from catch Quality indictors do not always agree on the spoilage rate of sample groups 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.
Partial Least Squares Regression (PLSR) model to 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] H 2 S sensor - T acc Sensory Quality [Y] Torry score NH 3 sensor TVB-N - ph CO sensor Pseudomonas spp. H 2 S producers P. phosphoreum TVC (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, H 2 S and NH 3 )
Partial Least Squares Regression (PLSR) model for haddock samples (0 C, 7 C, 15 C - 0 C, 0 C + abuse) Models based on pseudomonads counts and the CO, NH 3 H 2 S sensors, and T acc (time / temp) gave best results to predict sensory scores (r 2 =0.94; RMSEP: 0,49) PC2 TEMPERATURE High temp. ( 7-15 C) Low temp. (0 C) PC1 TIME 5% error between predicted sensory scores and experimental values when using a subset of the data
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) + abuse Process 2 Superchilling (CBC) Factory 2: Process 3 Traditional (cooling of fillets) Important to study natural products Catching, handling and storage conditions were not the same The aim was to select the best quality indicators (microbial, chemical, e-nose) to predict the sensory quality
Influence of different processes and temperature conditions (-1.5 C superchilling, 0.5 C and abuse) on the shelf-life of cod fillets Factory I Factory II Process 1 (ice) Process 2 (ice slurry) Process 3 (ice) Temperature C 10 9 8 7 6 5 4 3 2 1 0-1 -2 Traditional no cooling! 16 C 8h! 17 C 16h D 8d 0.5 C A 12 d! 16 C 8h B 11.5d CBC 0.5 C C 12.5 d E 14.5 d Temperature of fillets in experimental groups D A B C E G F I H Initial temperature Superchilling - CBC 0.5 C -1.5 C G 15 d F 15+d Traditional - cooling I 10.5 d H 13.5 d Average temperature of fillets! RT 16h 0.5 C 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Days from catch
Partial Least Squares Regression (PLSR) model to 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 sensor TVB-N Sensory Quality [Y] Torry score P.phosphoreum TVC pseudomonads H 2 S-producers 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, H 2 S-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 1 Temp. fluctuations: A, B, D, C Process 3 Cooling: H, I Process 2 Superchilling: E, F, G 85% correct classification based on sensory criteria (Torry score 7)
CONCLUSIONS Volatile compounds contributed by different SSOs give information about the spoilage processes in fish Electronic nose is a multi-indicator indicator device selective sensors for ketones,, amines, alcohols, aldehydes,, acids, esters and sulfur compounds e-nose can give more information than a single reference measurement for example TVB-N. Establishment of quality critera 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 storage conditions.
FUTURE PERSPECTIVES Implementation 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? sampling temperature control exclusion of background odors and contaminants sensitive / selective sensors for quality indicating compounds rapid detection of SSOs specific applications - establishment of quality criteria for different products and temperature conditions
Acknowledgements Funding: Icelandic Centre for Research AVS Research Fund of the Ministry of Fisheries Participating companies: Maritech Tros Tangi Skaginn (www.skaginn.is/)