nh

.K.gor,alay

orkromor igesuralye oPCAPCA

he puunfairt practthenticheapecially inancial

Meat Science 88 (2011) 638644

Contents lists available at ScienceDirect

Meat Sc

l scomminuted products have been a widespread problem in some retailmarkets, while meat species identication is a major global concern(Murugaiah et al., 2009). Identication of the species of origin in meatsamples is relevant to consumers for several reasons. The fallout fromfraudulent substitution or adulteration will possibly lead to economiclosses, jeopardize the health of consumers whomay have specic foodallergies and emotional disturbance due to religious reasons (Asensio,Gonzlez, Garca, & Martn, 2008; Bonne & Verbeke, 2008; Ghovvati,Nassiri, Mirhoseini, Moussavi, & Javadmanesh, 2009; Haunshi et al.,2009).

(2005) also successfully applied the FTIR spectroscopy technique todetect adulteration due to lard in cakes. The effectiveness of usingDNA-based technology such as polymerase chain reaction (PCR), forspecies identication in meat and fats has also been successfullycarried out by Aida, Che Man, Wong, Raha and Son (2005). The PCRmethod for species identication from pork and lard samples has beendemonstrated to be a potentially reliable technique for detection ofpig meat and fat. Chromatographic techniques such as HighPerformance Liquid Chromatography (HPLC) have also beenemployed in order to distinguish lard from other animal fatsThe advance in food technology has resulmore complicated where ingredients used into understand by the consumers unless theythe related eld. Additionally, the task of foo

Corresponding author. Tel.: +60 3 89430405; fax: +E-mail address: yaakobcm@gmail.com (Y.B. Che Ma

0309-1740/$ see front matter 2011 Elsevier Ltd. Aldoi:10.1016/j.meatsci.2011.02.022rewards for substitutionLai, Kemsley, & Wilson,at species in ground and

Mirghani (2001) developed a Fourier-transform infrared (FITR)spectroscopic method for detecting the presence of lard in certainanimal fat mixtures. Che Man, Syahariza, Mirghani, Jinap and Bakarof cheaper ingredients are relatively high (1995). Problems related to adulteration of me1. Introduction

Testing of food products for tauthentication is necessary to avoidconsumer protection against fraudulenOne of the major issues concerning auraw materials are substituted withKemsley, & Wilson, 1997) and espeadded products, where the potential rpose of labeling andcompetition and assureices in the food industry.city is where high valuer materials (Al-Jowder,n cases involving value

totally rely on the expertise of the practitioners of the related eldalone, but also require the contributions of those who are from othertechnical elds for example, chemistry and veterinary science. In halalauthentication, one cannot rely solely on physical inspection anddocumentation, but will need complementary evidence of the latesthigh technology analytical instrumentation. There aremany analyticaltechniques that have been successfully applied to detect and identifyadulteration of porcine based ingredients in foods. Che Man andted in the issues gettingfoods are more difcultare directly involved ind authentication cannot

(Marikkar, Ghazmethods mentiotime consumingto develop a rapthe Electronic noreliable method.of lard adulterainvestigated by C

60 3 89439745.n).

l rights reserved. 2011 Elsevier Ltd. All rights reserved.

Principal component analysis accounted by PC1.Rapid identication of pork for halal authechromatography mass spectrometer with

M. Nurjuliana a, Y.B. Che Man a,, D. Mat Hashim a, Aa Halal Products Research Institute, Universiti Putra Malaysia, 43400 UPM, Serdang, Selanb Department of Food Science, Faculty of Food Science and Technology, Universiti Putra M

a b s t r a c ta r t i c l e i n f o

Article history:Received 15 July 2010Received in revised form 31 January 2011Accepted 22 February 2011

Keywords:AuthenticationPorkElectronic noseSurface acoustic wave

The volatile compounds of pchromatography mass spectwas successfully employed fand chicken meats and sausafrom the frequency of theemployed to separate and ancompounds for the purposinterpretation. Analysis bysausages. It was shown that

j ourna l homepage: www.etication using the electronic nose and gaseadspace analyzer

S. Mohamed b

Malaysiasia, 43400 UPM, Serdang, Selangor, Malaysia

, other meats andmeat products were studied using an electronic nose and gaseter with headspace analyzer (GCMS-HS) for halal verication. The zNosedentication and differentiation of pork and pork sausages from beef, muttons which were achieved using a visual odor pattern called VaporPrint, derivedface acoustic wave (SAW) detector of the electronic nose. GCMS-HS wasze the headspace gasses from samples into peaks corresponding to individualf identication. Principal component analysis (PCA) was applied for datawas able to cluster and discriminate pork from other types of meats andcould provide a good separation of the samples with 67% of the total variance

ience

ev ie r.com/ locate /meatsc iali, Che Man, Peiris, & Lai, 2005). However, thened previously are generally sample destructive andin order to accomplish a complete analysis. The needid method for halal screening arose from the fact thatse can offer a non destructive, relatively low cost andThe potential use of the electronic nose for detectiontion in RBD palm olein has also been successfullyhe Man, Gan, NorAini, Nazimah and Tan (2005).

am of pork. SAW detector response for pork.

Table 2Major volatile components of pork by HSGCMS.

Peak Retention time (min) Compounds

1 0.461 Phenol2 2.295 Hexanal3 4.432 2-Butanone4 6.616 1-methoxyl-2-methyl-2-Pentanone5 11.821 Heptanal6 15.466 Benzaldehyde7 16.219 Heptyl ester 1-heptanol8 17.083 2-pentyl-furan9 17.674 Octanal10 18.671 1-Hexanol11 19.264 2, 4-Dimethylamphetamine12 19.739 2-Octenal13 20.190 1-Octyl-triuroacetate14 20.837 Butanal

639M. Nurjuliana et al. / Meat Science 88 (2011) 638644The aim of this study was to investigate the use of an electronicnose based on surface acoustic wave sensor for detection of pork andits discrimination from other types of meat and meat sausages. Foraroma proling and identication of the components that contributeto the avor of pork, gas chromatography mass spectrometer withheadspace analyzer (GCMS-HS) was employed.

2. Materials and methods

2.1. Meat samples

Meat samples from sheep, cow, chicken and pork were used in theexperiments. For sausages, two pork, one chicken and one beef sausageswere used. All samples were purchased from the local wet market inSerdang, Selangor,Malaysia. All sampleswere stored at20 C in ordertominimizeanydeteriorative changes to the samples. The sampleswerenot subjected to any pre treatment that may have altered their aroma

Fig. 1. Typical electronic nose chromatogrcomponents.

2.2. The electronic nose apparatus

The electronic nose (7100 vapor analysis system, Electronic SensorTechnology, New Bury Park, USA) is a portable bench top enclosure foreld laboratories or xed on-line installations. The zNose is baseduponwell knownprinciples of gas chromatography. This electronic noseuses a single, uncoated, high quartz surface acousticwave (SAW) sensorwhich consists of an uncoated 500 MHz acoustic interferometer orresonator bonded to a Peltier thermoelectric heat pumpwith the abilityto heat or cool the quartz crystal. This detector possesses advantages

Table 1Tentative identication of volatile compounds of rawpork from the electronic noseprole.

Peak Kovat's indices Compounds Odor description

1 612 Diacetyl Buttery2 718 3-hydroxy-2-butanone Buttery3 806 2-methyl-propanal Pungent4 903 Heptanal Fatty5 1000 Trimethyl pyrazine Roasted6 1104 Nonanal Soapy7 1207 Decanal Soapy

15 21.288 Nonanal16 21.971 Decyl ester17 22.909 2-Heptadecenal18 23.104 Pentasiloxane19 23.585 Naphtalene20 23.704 Acetic acid21 23.793 3-methyl-3,5 tetrahydro-4-thiopyranone22 23.858 Acetamide23 24.013 Dodecane24 24.185 Nitro-L-arginine25 25.615 2-Decenal26 26.410 2-Heptadecenal27 26.725 2H-Pyran28 26.986 2,4-Decadienal29 28.043 2-Undecenal30 28.844 Tetradecane31 29.052 2, 4-Bis(hydroxylamino)-6-methylpyramidine32 30.986 Pentadecane33 31.230 1-Heptadecanamine34 31.390 Sulfur35 32.755 Thiophene-3-ol36 33.010 Dodecane37 33.366 Hexadecanal38 34.500 Propenoic acid39 34.916 Sulfurous acid40 35.206 Hexadecanal41 35.984 3, 5-di-tert-Butyl-4-hyroxybenzaldehyde42 37.041 Octadecanal43 38.780 2-Octamine

were concentrated inaTenax trap (inlet 200 C) and carefully controlledto produce a repeatable and accurate collection of ambient vapors for

was performed using a headspace auto sampler (Model G1888, Agilent

rk s

640 M. Nurjuliana et al. / Meat Science 88 (2011) 638644analysis in the next step. In the second step, the trapwas rapidly heatedand the released vapors were re-focused at the head of a relatively lowsuch as high sensitivity, easy handling of signal, low power and longterm stability.

2.3. Electronic nose analysis

Five gramsof eachmeat samplewasminced andweighed into septa-sealed vials and prepared in triplicate using the same unit. After pre-cooking in a heated water bath maintained at 60 C for 10 min, thesample's vapor was pumped for 3 s into the electronic nosewith a side-ported sampling needle through the septa. The electronic nose analysisinvolved a two-step process. For the rst step, the vapors of the sample

Fig. 2. Typical electronic nose chromatogram of potemperature (40 C) capillary column (DB-5). This system is based onthe principle of gas chromatography. The column temperature wasprogrammed to heat from 40 to 160 C at a rate of 5 C/s, following alinear rise to its maximum temperature. This will cause the differentchemical component in the sample to be released, travel through thecolumn and land on the surface of the SAW. The SAW sensor wasoperated at a temperature of 30 C. When volatiles are adsorbed on the

Table 3Tentative identication of volatile compounds of pork sausage from the electronicnose prole.

Peak Kovat's indices Compounds Odor description

1 435 Ethanal Pungent2 646 2-methybutanal Pungent3 657 2-methyl pentan-3-one Mint4 709 Ethyl propionate 5 754 2-pentanone Ethereal6 820 Butanoic acid Rancid7 925 Hexanethiol Sulfur8 1043 Phenylmethanol 9 1140 -a-Prenchyl alcohol Camphor10 1212 5-methyl-2-furanaldehyde Almond11 1255 (t)-(4s)-carvone Caraway12 1902 Isoeugenol Floral13 2100 Octadecanoic acid Technologies, Palo Alto, CA, USA). The transfer line from the headspacesamplerwasdirectly connected to the injector of the gas chromatograph(GC). The oven was set at 110 C. The extraction conditions in theheadspace auto samplerwere programmed as follows: 20.0 min for vialsurface of the sensor, the frequencyof the SAWwill be alteredandwill inturn affect the detection signal and allow the identication of thecontaminants. The ow rate of puried heliumwas xed at 3.0 mL/min.The total cycle time per sample was 15 s.

2.4. Gas chromatography mass spectrometer with headspaceanalyzer analysis

Five grams of each meat sample was transferred into a 20 mlheadspace vial. The extraction of the volatile compounds of the samples

ausage. SAW detector response for pork sausage.equilibration, 0.20 min for vial pressurization, 0.20 min for lling theinjection loop, 0.05 min for loop equilibration and 1.0 min for sampleinjection. Helium with a purity of 99.999% was used for vialpressurization and as carrier gas. The volatile compoundswere analyzedusing a GC MS (Model 7890, Agilent Technologies, Palo Alto, CA, USA)equipped with a non polar column (J&W Scientic DB-5; 30 m, ID0.25 mm, lm thickness 0.25 m). The column temperature was kept at40 C for 10 min, increased at 6 C/min to 240 C and isothermallymaintained for 20 min. Themass selective detector (Model MSD 59556,Agilent Technologies, Palo Alto, CA, USA)wasused in electron ionizationmode. A mass range between 30 and 550 m/z was scanned. The massspectra obtained were compared to the NIST Mass Spectral SearchProgram for compound identication.

2.5. Statistical analysis

Unsupervised multivariate analysis, principal component analysis(PCA) was used for data processing using the Unsrambler v.9.6(CAMO AS, Trondheim, Norway) software. They were computediteratively in such a way that the rst principle component is the onethat carries the most information (or in statistical terms, the mostvariance explained). The second principle component will then carrythe maximum share of the residual information. Therefore, PCA ndsan alternative set of coordinate axis, principal components, of whichthe data set may be represented (He, Li, & Shao, 2007; Li & He, 2006;

Zhang et al., 2006). The two main aims of PCA are the reduction in thenumber of variables and elimination of redundancy.

3. Results and discussion

The electronic nose was used for rapid qualitative detection anddiscrimination of pork from other types of meat and meat productswhile the gas chromatography mass spectrometer with headspaceanalyzer (GCMS-HS) was used for aroma proling of pork and othermeats.

3.1. Volatile compounds of pork

The chromatographic proles of raw meat aroma of pork, beef,mutton and chicken obtained by electronic nose are shown in Fig. 1.The chromatogram from the electronic nose is a graphical display ofthe derivative of the frequency change versus time. Each peak foundin this derivative plot corresponds to a specic volatile compound andhas a retention time (given in seconds)which is specic to the columnand analysis temperature. The area under the peak was correlated tothe compound concentration andwas expressed in counts (cts). Therewere 7 (peaks: 1, 2, 3, 4, 5, 6 and 7) common compounds for all meatsamples. However, each sample showed variations in the amount for

every compound. Table 1 shows the peaks of the volatile compoundsand their odor description. The identication of the peaks wastentatively carried out based upon Kovat's indices database stored inthe substance library of the Microsense software using n-alkanes asstandards. In contrast to well known analytical instruments for theanalysis of avor compounds, the electronic nose does not give anyidentication of the compounds present whereas it attempts tointegrate measurements of the total headspace volatile compoundsand produce an aroma pattern that will exhibit differences orsimilarities among the samples (Arnold & Senter, 1998; Boothe &Arnold, 2002). Table 2 shows that there were a total of 43 volatilecomponents of pork identied by the GCMS-HS. The majority of thecompounds are well known lipid oxidation products, includingaldehydes, ketones and alcohols. It was observed from the aromaanalysis of pork using both techniques that the compounds werepositively identied and the most detected were aldehydes andketones. Most studies reported that almost all the aldehydes presentin pork such as heptanal and nonanal are oxidation products of oleicacid and linoleic acids which were the most abundant unsaturatedfatty acids of pork (Meinert, Andersen, Bredie, Bjergegaard, & Aaslyng,2007; Schiliemann, Wolm, Schrodter, & Ruttloff, 1987; Wettasinghe,Vasanthan, Temelli, & Swallow, 2001). From these results it can beshown that the electronic nose is a potentially feasible method for the

641M. Nurjuliana et al. / Meat Science 88 (2011) 638644Fig. 3. VaporPrint of different meats and sausages. 2 dimensional olfactory images which provide the odor concentration and characteristic shape for each sample.

Fig. 4. Four different meats (score plot) in principal component analysis of the electronic nose data. Abbreviations: p1, p2, p3, Pork1, Pork2, Pork3; c1, c2, c3, Chicken1, Chicken2,Chicken3; m1, m2, m3, Mutton1, Mutton2, Mutton3; b1, b2, b3, Beef1, Beef2, Beef3.

Fig. 3 (continued).

642 M. Nurjuliana et al. / Meat Science 88 (2011) 638644

Fig. 5. Four different meats and 3 different sausages (score plot) in principal component analysis of the electronic nose data. Abbreviations: p1, p2, p3 Pork1, Pork2, Pork3; c1, c2, c3Chicken1, Chicken2, Chicken3; m1, m2, m3 Mutton1, Mutton2, Mutton3; b1, b2, b3 Beef1, Beef2, Beef3; ps1, ps2, ps3, ps4, ps5 Pork sausage1, Pork sausage2, Pork sausage3, Porksausage4, Pork sausage5; cs1, cs2, cs3, cs4, cs5 Chicken sausage1, Chicken sausage2, Chicken sausage3, Chicken sausage4, Chicken sausage5; bs1, bs2, bs3, bs4, bs5 Beef sausage1,Beef sausage2, Beef sausage3, Beef sausage4, Beef sausage5.

643M. Nurjuliana et al. / Meat Science 88 (2011) 638644analysis of aroma of rawmaterials and can possibly offer a technologywhich can rival GCMS-HS.

The gas chromatograms of the aroma ofmeat products are shown inFig. 2. The chromatograms show that pork sausage contained morearomatic compounds and this observation is based on the number ofpeaks. Table 3 shows the set of volatile compounds corresponding topeaks 113 and their odor descriptions while Fig. 3 displays the aromapattern of 4 different types ofmeat, namely pork (Fig. 3a), chickenmeat(Fig. 3b), mutton (Fig. 3c) and beef (Fig. 3d) and 3 different types ofsausages, namely, pork sausage (Fig. 3e), chicken sausage (Fig.3f) andbeef sausage (Fig. 3g). The VaporPrint is a 2-dimensional olfactoryimagewhich provides the odor concentration and characteristic shapes(Sim et al., 2003). In this polar format, the display starts at the 0.0position and follows around the dialwith retention times increasing in aclockwise direction. By measuring the time required for each chemical

to reach the sensor and the amount affects the SAW crystal's vibration

Fig. 6. Seven electronic nose variables (loading plot) in pand hence both the identity (retention time) and the quantity (amount)of the substance canbe calculatedusing the software incorporated in thedevice. Different meat and sausage samples showed variations in theamounts of each compound and generate a unique and easilyrecognizable image. The unique nature of this display is subjected tothe relative concentrations of several compounds making up the mixwhile alsoquantifying the strengthof each chemical compoundwithin asample. However, to get an overall view of the complex data, principalcomponent analysis was carried out (Figs. 4 and 5).

3.2. Principal component analysis

Principal component analysis (PCA) as an unsupervised classica-tion method to visualize the resemblance and difference amongdifferent measurements in the data sets was used in order to structure

the data matrix. The meat samples were separated along the rst PC

rincipal component analysis of the electronic data.

which described 67% of the peak variations (Fig. 5) and showed sevendened groups. Along the PC1 axis, pork, chicken, mutton, beef andbeef sausages were located with high positive scores but on the otherhand along the PC3 axis, chicken sausage had low positive scoreswhile pork sausage had low negative scores. This percentage appearsto sufciently dene a goodmodel, especially for qualitative purposes.Fig. 6 shows the loading plot with seven variables. Only four variables(2, 4, 6 and 7) had a far Euclidean distance from the origin while theremaining variables were considered as unimportant for discrimina-tion (low loading values along PC1 and close to the origin). The highpositive correlation between peak 4 and PC1 indicated that thevolatile prole of pork contained a higher proportion of heptanal(peak number 4) (Table 1). This indicates that the heptanal has amajor inuence upon the discrimination of pork from other types ofmeats and sausages. This concurred with the observation of Shahidi(1994) who reported that aldehydes are the major componentsidentied in the volatiles of cooked pork.

Al-Jowder, O., Kemsley, E. K., & Wilson, R. H. (1997). Mid-infrared spectroscopy andauthenticity problem in selected meats: A feasibility study. Food Chemistry, 59,195201.

Arnold, J. W., & Senter, S. D. (1998). Use of digital aroma technology and SPME GCMSto compare volatile compounds produced by bacteria isolated from processedpoultry. Journal of the Science of Food and Agriculture, 78(3), 343348.

Asensio, L., Gonzlez, I., Garca, T., & Martn, R. (2008). Determination of food authenticityby enzyme-linked immunosorbent assay (ELISA). Food Control, 19(1), 18.

Bonne, K., & Verbeke, W. (2008). Muslim consumer trust in halal meat status andcontrol in Belgium. Meat Science, 79(1), 113123.

Boothe, D. D.., & Arnold, J. W. (2002). Electronic nose analysis of volatile compoundsfrom poultry meat samples, fresh and after refrigerated storage. Journal of theScience of Food and Agriculture, 82(3), 315.

Che Man, Y. B., Gan, H. L., NorAini, I., Nazimah, S. A. H., & Tan, C. P. (2005). Detection oflard adulteration in RBD palm olein using an electronic nose. Food Chemistry, 90(4),829835.

Che Man, Y., & Mirghani, M. (2001). Detection of lard mixed with body fats of chicken,lamb, and cow by fourier transform infrared spectroscopy. Journal of the AmericanOil Chemists' Society, 78(7), 753761.

Che Man, Y. B., Syahariza, Z. A., Mirghani, M. E. S., Jinap, S., & Bakar, J. (2005). Analysis ofpotential lard adulteration in chocolate and chocolate products using Fouriertransform infrared spectroscopy. Food Chemistry, 90(4), 815819.

Ghovvati, S., Nassiri, M. R., Mirhoseini, S. Z., Moussavi, A. H., & Javadmanesh, A. (2009).Fraud identication in industrial meat products by multiplex PCR assay. FoodControl, 20(8), 696699.

Haunshi, S., Basumatary, R., Girish, P. S., Doley, S., Bardoloi, R. K., & Kumar, A. (2009).

644 M. Nurjuliana et al. / Meat Science 88 (2011) 638644The ability of the zNose to qualitatively discriminate and clusteramong 4 common meat samples and 3 types of sausage wasdemonstrated in this study. Measurements of the volatile compoundsby GCMS-HS were also employed which indicate that the electronicnose has adequate selectivity and sensitivity to perform avordetection in meats. With a total analysis of less than a minute andrequiring less than 5 g of sample, the electronic nose offers a rapid,accurate, low cost and environmentally friendly tool for detection ofporcine based ingredients in foods and this is especially useful forhalal authentication and verication.

Acknowledgment

This research work was supported by Universiti Putra Malaysia(Grant No. Research University Grant Scheme: 91033) awarded toProfessor Dr. Yaakob Bin Che Man. The authors are also greatlyindebted to Mr. Tibby Lim for his technical support, Dr. Marina AbdulManaf, Miss Syahariza Zainul Abidin and Mdm. Siti Munira AbdukRazak for their assistance.

References

Aida, A. A., Che Man, Y. B., Wong, C. M. V. L., Raha, A. R., & Son, R. (2005). Analysis of rawmeats and fats of pigs using polymerase chain reaction for Halal authentication.Meat Science, 69(1), 4752.Identication of chicken, duck, pigeon and pig meat by species-specic markers ofmitochondrial origin. Meat Science, 83(3), 454459.

He, Y., Li, X., & Shao, Y. (2007). Fast discrimination of apple varieties using Vis/NIRspectroscopy. International Journal of Food Properties, 10(1), 918.

Lai, Y. W., Kemsley, E. K., & Wilson, R. H. (1995). Quantitative analysis of potentialadulterants of extra virgin olive oil using infrared spectroscopy. Food Chemistry, 53(1), 9598.

Li, X., & He, Y. (2006). A novel approach to pattern recognition based on PCA-ANN inspectroscopy. Advanced data mining and applications (pp. 525532).

Marikkar, J. M. N., Ghazali, H. M., Che Man, Y. B., Peiris, T. S. G., & Lai, O. M. (2005).Distinguishing lard from other animal fats in admixtures of some vegetable oilsusing liquid chromatographic data coupled with multivariate data analysis. FoodChemistry, 91(1), 514.

Meinert, L., Andersen, L. T., Bredie, W. L. P., Bjergegaard, C., & Aaslyng, M. D. (2007).Chemical and sensory characterisation of pan-fried pork avour: Interactionsbetween raw meat quality, ageing and frying temperature. Meat Science, 75(2),229242.

Murugaiah, C., Noor, Z. M., Mastakim, M., Bilung, L. M., Selamat, J., & Radu, S. (2009).Meat species identication and Halal authentication analysis using mitochondrialDNA. Meat Science, 83(1), 5761.

Schiliemann, J., Wolm, G., Schrodter, R., & Ruttloff, H. (1987). Chicken avor-formation,composition, and production. Part 1: Flavor precursors. Nahrung, 31, 4756.

Shahidi, F. (1994). Flavor of meat and meat products and overview. In C. A. Hall (Ed.),Flavor of meat and meat products (pp. 13).

Sim, C., Ahmad, M., Ismail, Z., Othman, A., Noor, N., & Zaihidee, E. (2003). Chemometricclassication of herba Orthosiphon stamineus according to its geographical originusing virtual chemical sensor based upon fast GC. Sensors, 3(10), 458471.

Wettasinghe, M., Vasanthan, T., Temelli, F., & Swallow, K. (2001). Volatile avourcomposition of cooked by-product blends of chicken, beef and pork: A quantitativeGCMS investigation. Food Research International, 34(23), 149158.

Zhang, Q., Zhang, S., Xie, C., Zeng, D., Fan, C., Li, D., et al. (2006). Characterization ofChinese vinegars by electronic nose. Sensors and Actuators B: Chemical, 119(2),538546.4. Conclusion

Rapid identification of pork for halal authentication using the electronic nose and gas chromatography mass spectrometer with headspace analyzerIntroductionMaterials and methodsMeat samplesThe electronic nose apparatusElectronic nose analysisGas chromatography mass spectrometer with headspace analyzer analysisStatistical analysis

Results and discussionVolatile compounds of porkPrincipal component analysis

ConclusionAcknowledgmentReferences