falsificari2.pdf

8/12/2019 falsificari2.pdf

1/8


2/8


3/8

Eur Food Res Technol

1 3

pass through filter paper, and the filtrate was placed in a

water bath at a temperature of 70 C and purged with N2

gas to accelerate the evaporation of hexane. After comple-

tion of the evaporation, obtained unary fat samples were

stored under refrigeration conditions until the Raman

measurements.

Instruments and data collection

Raman measurements were performed using a DeltaNu

Examiner Raman Microscopy system (DeltaNu Inc., Lara-

mie, WY, USA) with a 785-nm laser source and a cooled

charge-coupled (CCD, at 0 C) detector. The fat sam-

ples were kept in water bath for one hour at 70 C before

Raman measurements. Spectra were obtained in the range

of 2002,000 cm1at a resolution of 2 cm1. The full spec-

trum of Raman data was used for the presentation where

a gap between each spectrum is created manually by add-

ing a certain value on top of the initial intensity value of

the each spectrum to prevent the overlapping of the spectrafor better illustration. Experimental parameters were as fol-

lows: laser power level of 100 mW and integration time of

5 s for fish samples and 30 s for the rest. All measurements

were conducted in duplicate for each sample.

Data processing and chemometric analysis

The application of chemometric methods in Raman spec-

troscopic studies was newly developed due to difficul-

ties in the analysis of abundant quantity of information.

In this study, the collected Raman data were used to cre-

ate principal component analysis (PCA) models. PCA was

performed using stand-alone chemometrics software (Ver-

sion Solo 6.5 for Windows 7, Eigenvector Research Inc.,

Wenatchee, WA, USA) and the singular value decomposi-

tion method.

Before forming the model, the data that were collected

from Raman measurements were preprocessed by baseline

correction, different orders of derivatives, mean center, nor-

malizing, smoothing and auto-scaling to achieve succeed-

ing models. All derivative computations were executed by

the Savitsky-Golay convolution using a second-order pol-

ynomial. Each sample was exposed to different combina-

tions of preprocessing techniques; then, the combination

of technique that provided with the best classification ratio

was chosen. PCA was performed in four stages to develop

different PCA models. The developed PCA models were

PCA model 1, first derivative, mean center and smoothing;

PCA model 2, fourth derivative; PCA model 3, third deriv-

ative and PCA model 4, second derivative. Principal com-

ponent (PC) scores with the highest percentage of variance

were used to plot the score graphs for the evaluation of the

success of the model. In stage 1, it was aimed to create as

many groups as possible which comprised of the fat sam-

ples from each species. Although, the raw Raman data col-

lected for all the species were used for the developments of

all the models, only the related PC scores of samples whose

separation was desired were assessed for the presentation

of each step.

Results and Discussion

Raman spectra of the fat samples

Raman spectra of seven different meat species investigated

in this study are shown in Fig. 1. The most intense band for

the Raman spectra of all the species is situated around 868

872 cm1. The other less intense bands around 1,0811,085

and 1,3031,306 cm1followed this strong band. The band

around 1,4431,451 cm1 has the smallest Raman inten-

sity for all the species. Explanations of the most noticeable

Raman bands of the fat and salami samples are as follows.The bands around 800 to 920 cm1 region were

explained by (C1-C2) stretch of the acyl chain and end

methyl rocking mode of the all-trans chains [52]. The bands

positioned around 1,070 to 1,110 cm1have been assigned

to (CC) stretching modes in the gauche conformation.

Bands located between 1,295 and 1,305 cm1 correspond

to (CH2) methylene twisting deformations [53]. Bands in

the region of 1,400 to 1,500 cm1have been assigned to

(CH2) methylene scissor deformations [54].

In order to prevent regional differences in one species, it

was a prerequisite to obtain almost identical Raman spec-

tra from different parts of the same animal. Our results met

with this prerequisite such that fat samples obtained from

seven different parts of cattle namely perirenal, abdominal,

brisket, heart, foreleg, steak and thigh had identical Raman

Fig. 1 Raman spectra of extracted fat samples of the meat species.

The spectra are offset for clarity by 12,000 unit


4/8


1 3

spectra as shown in Fig. 1S, and this was confirmed for

sheep samples (brisket, foreleg, tail, chop and perirenal)

as shown in Fig. 2S. The results obtained for different fish

species (trout, sardines and salmon) were also in agreement

with this as shown in Fig. 3S.

Data analysis

Data analysis of the fat samples

Data processing of the Raman spectra was performed by

the usage of developed PCA models. In order to utilize

the spectral variations between meat species, PCA was

applied on the whole data ranging from 200 to 2,000 cm1.

Analysis of the collected Raman data was carried out in

a total of four stages. The first stage was the classifica-

tion of meat species. First two principal components (PC1

and PC2) with the highest percentage of variance (65.86

and 20.98 %, respectively) were chosen, and the principal

component scatter plot of PC1 vs. PC2 was constructed(Fig. 2). As seen in Fig. 2a, the four clearly separated

clusters were found in the areas of A1, A2, A3 and A4; in

each area, different species were included. Area A1 com-

prises cattle, sheep and goat samples; area A2 includes pig

and buffalo samples; area A3 consists of poultry samples

and finally fish samples are convened in area A4. Applied

preprocessing techniques in the first stage (PCA model 1)

were sufficed for classification of poultry and fish species

in the A3 and A4 areas. However, in order to eliminate the

interferences in the areas of A1 and A2, several preproc-

essing techniques were individually tested to increase the

performance of decomposition. As seen in Fig. 2a, area

A1 contains the samples of three different meat species

(cattle, sheep and goat). Thus, subsequent PCA models

(2, 3 and 4) were developed to separate the group of cat-

tle, sheep, goat, buffalo and pig species. The separation of

goat samples from the clusters of cattle and sheep was the

purpose of the second stage. In PCA model 2, a different

preprocessing technique, which involves taking the fourth

order of derivative, was found as the best way to separate

the goat samples. The variance of 85.07 and 6.41 % was

explained by PC1 and PC2, respectively. As shown in

Fig. 2b, goat samples were successfully separated from

the cattle and sheep clusters. In PCA model 3, the cattle

and sheep samples were situated in different regions of the

score plot by applying the third order of derivative to the

raw Raman data (Fig. 2c) and A1 was divided into three

different zones called A1 Goat (A1G), A1 Cattle (A1C) and

A1 Sheep (A1S). PC1 and PC2 explained 97.21 and 1.81 %

of the variance, respectively. In the fourth stage, pig and

buffalo clusters were separated from each other by appli-

cation of the second derivative (Fig. 2d), and the area of

A2 was converted into two different zones called A2 Pig

(A2P) and A2 Buffalo (A2B). Finally, all PC scores for

each species assigned by the PCA models are denotated in

a single figure as shown in Fig. 2e. This figure was only

a demonstration in which the seven different species were

successfully discriminated from each other. The clusters of

the poultryfish, goat, cattlesheep and pigbuffalo species

were separated from first, second, third and fourth stages,

respectively.All stages of the developed analysis approach are sum-

marized in Table 1. When the developed system is con-

fronted with an unknown sample, each stage of the analy-

sis will be applied in the order of above-mentioned table to

determine the origin of the sample.

Data analysis of salami samples:

The Raman spectra of the fat samples of cattle, pig and

chicken and their salami products are shown in Fig. 3.

The Raman spectra of extracted fat samples of buffalo,

sheep, chicken, goat and their salami products are shown inFig. 4S. Raw Raman data of extracted fat samples of meat

species and their salami product were used to develop an

analysis method for the discrimination of the samples. In

PCA model 1, first derivative, mean center and smoothing

were applied as preprocessing techniques. Data analysis

for salami products was successful, as long as the Raman

spectrum of a unary salami sample was located in the same

area with its original species. Additional preprocessing

techniques were required to resolve the interferences aris-

ing from cattlepig salami and cattlesheep salami as they

tend to be in the same cluster with their unary fat samples.

Application of a differential analysis in the fourth order

was the next step to separate the cattlepig salami samples

from the cluster of unary cattle fat samples. The first two

principal components, which had the highest percentage of

variance, were chosen, and PC score plot for extracted fat

samples of meat species and their salami products except

for fish was constructed (Fig. 4). PC1 and PC2 explained

85.07 % and 6.41 % of the variance, respectively. Third

derivative was applied to obtain the individual cluster of

cattlesheep salami samples, and PC1 and PC2 explained

the variance of 96.26 % and 2.15 %, respectively. Although

PC scores of the salami samples (binary mixtures) were in

the area of their original species with a certain distance, it

was not possible to analyze the salami samples composi-

tion quantitatively by looking at this distance. A three-

dimensional illustration of Fig. 4is shown in Fig. 5S.

As we know that analyzed salami samples were prepared

with the ratio of 5050 % (w/w), it is possible to interpret

the species in the salami sample by looking at the distance

of the sample to the clusters of its two original species.

Increasing the number of binary salami samples and ana-

lyzing salami samples containing different ratios of fats


5/8


1 3

(a) (b)

(c) (d)

(e)

Fig. 2 Stages of PCA: a classification of meat species; A1 (cattle,

sheep and goat), A2 (buffalo and pig), A3 (chicken and turkey), and

A4 (fish); b separation of goat samples from the cattle and sheep

samples (A1 (cattle and sheep): A1 (CS), A1 (Goat): A1G); csepa-

ration of cattle and sheep samples (A1C, A1S); d separation of pig

and buffalo samples (A2 (pig): A2P, A2 (buffalo): A2B); e denota-

tion of the results as a combination of all stages of the analysis; A1C,

A1G, A1S, A2B, A2P, A3, A4


6/8


1 3

will help us to determine the origin of an unknown sample

by giving a meaning to this distance.

Conclusions

Effectuality of the Raman spectroscopy was demonstrated

by this study as it offers a simple alternative for differentia-

tion of the meat species and determination of the origin of

meat products. Taking the advantages of both Raman spec-

troscopy and chemometric methods, successful classification

of seven different meat species and their salami products was

reported. Using fat samples extracted from meat for Raman

measurements is a simpler and more rapid way to explain

the obtained spectra, which enables us to eliminate the inter-

ferences arising from meat matrix. In this technique, a sig-

nificant reduction in analysis time was provided by Raman

measurements taking only 30 s and PCA which substan-

tially simplifies data processing. This promising methodol-

ogy could be useful in the detection of adulterations in meat

industry and contribute to pacifying consumers concernsabout the meat they consume. Increasing the number of meat

and meat product samples and application of different chem-

ometric methods will be investigated in our further studies as

they may improve the success of the developed method.

Supplementary Data Available

Supplementary data includes Raman spectra of the

extracted fat of meat species and their salami product sam-

ples, three dimensional illustration of the PC score plot

for all the meat species and their salami products, detailed

information about the origin of the extracted fat samples

and recipe for salami production is available.

Conflict of interest None.

Compliance with Ethics Requirements This article does not con-

tain any studies with human or animal subjects.

References

1. Barai BK, Nayak R, Singhal RS, Kulkarni PR (1992) Approachesto the detection of meat adulteration. Trends Food Sci Technol

3:6972

2. Sugarman C (1987) U. S. biologist spots counterfeit meat. 30:34

3. Ballin NZ, Vogensen FK, Karlsson AH (2009) Species determi-

nation - Can we detect and quantify meat adulteration? Meat Sci

83(2):165174

4. Vallejo-Cordoba B, Gonzalez-Cordova AF, Mazorra-Manzano

MA, Rodriguez-Ramirez R (2005) Capillary electrophoresis for

the analysis of meat authenticity. J Sep Sci 28(910):826836

5. Ashoor SH, Monte WC, Stiles PG (1988) Liquid-Chromato-

graphic Identification of Meats. Journal of the Association of

Official Analytical Chemists 71(2):397403

Table 1 General illustration of

cluster analysis of meat species

A1 (Ggoat, Ccattle, Ssheep);

A2 (Ppig,Bbuffalo); A3 (C

chicken, Tturkey); A4 (Ffish)

Stage areas (A)

First stage A1 (G-CS) A2 (PB) A3 (CT) A4F

Second stage A1G A1(CS) A2 (PB) A3 (CT) A4F

Third stage A1G A1C A1S A2 (PB) A3 (CT) A4F

Fourth stage A1G A1C A1S A2P A2B A3 (CT) A4F

Fig. 3 Raman spectra of the salami samples. The spectra are offset

for clarity by 13,500 unit

Fig. 4 PC score plot of the extracted fat samples of meat and salami

products


7/8


1 3

6. Toorop RM, Murch SJ, Ball RO (1997) Methodology and devel-

opment of prediction equations for the determination of pork sub-

stitution in veal. Food Res Int 30(8):629636

7. Chen F-C, Hsieh Y-H (2000) P.: A monoclonal antibody-based

ELISA for detection of pork in heat-processed meat products. J

AOAC Int 83(1):7985

8. Giovannacci I, Guizard C, Carlier M, Duval V, Martin JL,

Demeulemester C (2004) Species identification of meat products

by ELISA. Int J Food Sci Technol 39(8):863867

9. Hird H, Goodier R, Hill M (2003) Rapid detection of chicken and

turkey in heated meat products using the polymerase chain reac-

tion followed by amplicon visualisation with vistra green. Meat

Sci 65(3):11171123

10. Asensio L, Gonzalez I, Garcia T, Martin R (2008) Determina-

tion of food authenticity by enzyme-linked immunosorbent assay

(ELISA). Food Control 19(1):18

11. Ansfield M, Reaney SD, Jackman R (2000) Production of a sensi-

tive immunoassay for detection of ruminant and porcine proteins,

heated to > 130 degrees C at 2.7 bar, in compound animal feed-

stuffs. Food and Agricultural Immunology 12(4):273284

12. Ansfield M, Reaney SD, Jackman R (2000) Performance assess-

ment and validation of a sensitive immunoassay for detection of

ruminant and porcine heat stable proteins in compound animal

feedstuffs. Food and Agricultural Immunology 12(4):285297

13. Berger RG, Mageau RP, Schwab B, Johnston RW (1988) Detec-

tion of Poultry and Pork in Cooked and Canned Meat Foods by

Enzyme-Linked Immunosorbent Assays. Journal of the Associa-

tion of Official Analytical Chemists 71(2):406409

14. Destaillats F, Sebedio JL, Berdeaux O, Juaneda P, Angers P

(2005) Gas chromatography-mass spectrometry determination of

metabolites of conjugated cis-9, trans-11, cis-15 18: 3 fatty acid.

J Chromatogr B 820(1):1522

15. Quehenberger O, Armando AM, Dennis EA (2011) High sensi-

tivity quantitative lipidomics analysis of fatty acids in biological

samples by gas chromatography-mass spectrometry. Bba-Mol

Cell Biol L 1811(11):648656

16. Dayhuff LE, Wells MJM (2005) Identification of fatty acids in

fishes collected from the Ohio River using gas chromatography-

mass spectrometry in chemical ionization and electron impactmodes. J Chromatogr A 1098(12):144149

17. Abu EO, Oluwatowoju I (2009) Omega-3 index determined by

gas chromatography with electron impact mass spectrometry.

Prostag Leukotr Ess 80(4):189194

18. Hartig C (2008) Rapid identification of fatty acid methyl esters

using a multidimensional gas chromatography-mass spectrometry

database. J Chromatogr A 1177(1):159169

19. Woo AH, Lindsay RC (1980) Method for the Routine Quantita-

tive Gas-Chromatographic Analysis of Major Free Fatty-Acids in

Butter and Cream. J Dairy Sci 63(7):10581064

20. May WE, Hume DJ (1993) An Automated Gas-Liquid-Chroma-

tographic Method of Measuring Free Fatty-Acids in Canola. J

Am Oil Chem Soc 70(3):229233

21. Zhao GH, Liu JF, Nyman M, Jonsson JA (2007) Determination of

short-chain fatty acids in serum by hollow fiber supported liquidmembrane extraction coupled with gas chromatography. J Chro-

matogr B 846(12):202208

22. Coltro WKT, Ferreira MMC, Macedo FAF, Oliveira CC, Visen-

tainer JV, Souza NE, Matsushita M (2005) Correlation of animal

diet and fatty acid content in young goat meat by gas chromatog-

raphy and chemometrics. Meat Sci 71(2):358363

23. Rezanka T (2000) Analysis of very long chain polyunsaturated

fatty acids using high-performance liquid chromatography -

atmospheric pressure chemical ionization mass spectrometry.

Biochem Syst Ecol 28(9):847856

24. Boselli E, Pacetti D, Curzi F, Frega NG (2008) Determination of

phospholipid molecular species in pork meat by high performance

liquid chromatography-tandem mass spectrometry and evapora-

tive light scattering detection. Meat Sci 78(3):305313

25. Colombo F, Viacava R, Sacchelli D, Colombo M, Camisasca S

(1998) Differentiation of the Ostrich (Struthio camelus) and emu

(Dromaius novaehollandiae) species by Polymerase Chain Reac-

tion - Part I: PCR on the muscular tissue of Struthio camelus.

Industrie Alimentari 37(370):618619

26. Zhang GL, Zheng MG, Zhou ZJ, Ouyang HS, Lu Q (1999)

Establishment and application of a polymerase chain reaction for

the identification of beef. Meat Sci 51(3):23323627. Lockley AK, Bardsley RG (2000) DNA-based methods for food

authentication. Trends Food Sci Technol 11(2):6777

28. Chung ER, Kim WT, Kim YS, Han SK (2000) Identification of

Hanwoo meat using PCR-RFLP marker of MC1R gene associ-

ated with bovine coat color. Journal of Animal Science Technol-

ogy 42:379380

29. Hopwood AJ, Fairbrother KS, Lockley AK, Bardsley RG

(1999) An actin gene-related polymerase chain reaction (PCR)

test for identification of chicken in meat mixtures. Meat Sci

53(4):227231

30. Ganai TAS, Singh R (2000) K., Butchaiah, G. : DNA amplifica-

tion fingerprinting of cattle and buffalo genome by RAPD-PCR

utilizing arbitrary oligonucleotide primers. Buffalo Journal

3:331339

31. Koh MC, Lim CH, Chua SB, Chew ST, Phang STW (1998)

Random amplified polymorphic DNA (RAPD) fingerprints for

identification of red meat animal species. Meat Sci 48(34):

275285

32. Chikuni K, Ozutsumi K, Koishikawa T, Kato S (1990) Species

Identification of Cooked Meats by DNA Hybridization Assay.

Meat Sci 27(2):119128

33. Matsunaga T, Chikuni K, Tanebe R, Muroya S, Shibata K, Yam-

ada J, Shinmura Y (1999) A quick and simple method for the

identification of meat species and meat products by PCR assay.

Meat Sci 51:143148

34. Colombo F, Marchisio E, Pizzini A, Cantoni C (2002) Identifi-

cation of the goose species (Anser anser) in Italian Mortara

salami by DNA sequencing and a Polymerase Chain Reaction

with an original primer pair. Meat Sci 61(3):291294

35. Girish PS, Anjaneyulu A (2005) S. R., Viswas, K. N., Shivaku-

mar, B. M., Anand, M., Patel, M., Sharma, B. : Meat species

identification by polymerase chain reaction-restriction fragment

length polymorphism (PCR-RFLP) of mitochondrial 12S rRNA

gene. Meat Sci 70:107112

36. Downey G (1996) Authentication of food and food ingredients by

near infrared spectroscopy. J near Infrared Spec 4:4761

37. Cozzolino D, Murray I (2004) Identification of animal meat mus-

cles by visible and near infrared reflectance spectroscopy. Leb-

ensm-Wiss Technol 37(4):447452

38. Zhang HZ, Lee TC (1997) Rapid near-infrared spectroscopic

method for the determination of free fatty acid in fish and

its application in fish quality assessment. J Agr Food Chem

45(9):35153521

39. Arnalds T, McElhinney J, Fearn T, Downey G (2004) A hierar-chical discriminant analysis for species identification in raw meat

by visible and near infrared spectroscopy. J near Infrared Spec

12(3):183188

40. Fumiere O, Sinnaeve G, Dardenne P (2000) Attempted authen-

tication of cut pieces of chicken meat from certified production

using near infrared spectroscopy. J near Infrared Spec 8(1):2734

41. Baeten V, Hourant P, Morales MT, Aparicio R (1998) Oil and

fat classification by FT-Raman spectroscopy. J Agr Food Chem

46(7):26382646

42. Abbas O, Pierna JAF, Codony R, von Holst C (2009) Assessment

of the discrimination of animal fat by FT-Raman spectroscopy. J

Mol Struct 92426:294300


8/8


1 3

43. Yang H, Irudayaraj J, Paradkar MM (2005) Discriminant analysis

of edible oils and fats by FTIR, FT-NIR and FT-Raman spectros-

copy. Food Chem 93(1):2532

44. Baeten V, Meurens M, Morales MT, Aparicio R (1996) Detection

of virgin olive oil adulteration by Fourier transform Raman spec-

troscopy. J Agric Food Chem 44(8):22252230

45. El-Abassy RM, Donfack P, Materny A (2009) Visible Raman

spectroscopy for the discrimination of olive oils from different

vegetable oils and the detection of adulteration. J Raman Spec-

trosc 40(9):12841289

46. Marigheto NA, Kemsley EK, Defernez M, Wilson RH (1998) A

comparison of mid-infrared and Raman spectroscopies for the

authentication of edible oils. J Am Oil Chem Soc 75(8):987992

47. Baeten V, Aparicio R (2000) Edible oils and fats authentication

by Fourier transform Raman spectrometry. Biotechnol Agron Soc

Environ 4:196203

48. Sadeghijorabchi H, Wilson RH, Belton PS, Edwardswebb JD,

Coxon DT (1991) Quantitative-analysis of oils and fats by

Fourier-transform Raman-spectroscopy. Spectrochim Acta A

47(910):14491458

49. Beattie JR, Bell SEJ, Borgaard C, Fearon A, Moss BW (2006)

Prediction of adipose tissue composition using Raman spec-

troscopy: Average properties and individual fatty acids. Lipids

41(3):287294

50. Farhad SFU, Abedin KM, Islam MR, Talukder AI, Haider AFMY

(2009) Determination of ratio of unsaturated to total fatty acids in

edible oils by laser Raman spectroscopy. J Appl Sci 9(8):15381543

51. Aktas N, Genccelep H (2006) Effect of starch type and its modi-

fications on physicochemical properties of bologna-type sausage

produced with sheep tail fat. Meat Sci 74(2):404408

52. Kint S, Wermer PH, Scherer JR (1992) Raman-Spectra of

Hydrated Phospholipid-Bilayers. 2. Water and Headgroup Inter-

actions. J Phys Chem-Us 96(1):446452

53. Butler M, Salem N, Hoss W, Spoonhower J (1979) Raman Spec-

tral Analysis of the 1300 Cm-1 Region for Lipid and Membrane

Studies. Chem Phys Lipids 24(1):99102

54. Sadeghijorabchi H, Hendra PJ, Wilson RH, Belton PS (1990)

Determination of the total unsaturation in oils and margarines

by Fourier-transform Raman-spectroscopy. J Am Oil Chem Soc

67(8):483486

falsificari2.pdf

Documents