Food Control 48 (2015) 123e129

Contents lists available at ScienceDirect

Food Control

journal h o m e p age: www .elsevier.co m /lo c ate/foodcont

Discriminating organic and conventional foods by analysis of their microbial ecology: An application on fruits

Cline Bigot*, Jean-Christophe Meile, Andr Kapitan, Didier Montet

CIRAD-UMR Qualisud, TA B-95/16, 73, rue Jean-Franois Breton, 34398 Montpellier Cedex 5, France

a r t i c l e i n f o

Article history:Received 2 December 2013Received in revised form19 March 2014Accepted 22 March 2014Available online 2 April 2014

Keywords:Biological bare code Food microbiology Organic foodsPCR-DGGE TraceabilityMicrobial ecology

a b s t r a c t

Traceability of foods is mainly done at the administrative level, and the use of analytical tools is rare. Previous studies have demonstrated that microbial ecology analyses at the molecular level (such as PCR- DGGE) could be used to provide food with a unique biological signature that could be linked to the geographical origin of food. The present study aimed at testing this approach to differentiate farming types by analyzing organic and conventional food products. To this end, the microbial ecology of organic and conventional nectarines was analyzed and statistically compared.Our results show that yeast and bacterial communities were specic of the farming type allowing organic fruits to be discriminated from conventional ones. Several microbial species were identied as potential, biological markers which detection could be used to certify the origin as well as the mode of production of foodstuff. We proposed this analytical tool as a rst step to control and authentify organic foods. 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Following various food crises such as the mad cow disease or the recent fraud in the beef meat market, European consumers are more and more perceptible to the quality and the origin of food- stuffs they buy, and food safety became one of their main concerns (Lairon, 2010). As a response to these safety, sociologic and eco- nomic problems and within the framework of the globalization, the European regulation relative to the sanitary quality of foodstuffs had to be strengthened. The Food Law (European regulation CE No.178/2002), applied on the 1st of January 2005, imposes to all food- processing companies of the European Union (EU) to keep con- sumers informed about the nature of the product and any sanitary problems. Moreover, it imposes the traceability of foodstuffs at all steps of the food production. This regulation applies also to organic food industry.Organic farming is a method of sustainable production which contributes to the environmental and animal protection by a set of specic agricultural practices. According to the French Agency for Food Safety (Anses, http://www.anses.fr), organic farming is char- acterized by the use of a positive list of chemicals and Genetically

* Corresponding author. Tel.: 33 4 67 61 57 28; fax: 33 4 67 61 44 33.E-mail address: celine.bigot@cirad.fr (C. Bigot).

Modied Organisms are prohibited. This mode of production is supervised by a European regulation No. 834/2007 which denes the principles of production, preparation and importation to be respected, the lists of usable products, the practices for every type of breeding and the principles of control, certication, penalty and labelling. The organic farming mention is obtained after a period of land conversion of two or three years and a variable period of animal conversion according to species. During this period, organic farmers respect a rigorous specication which favours the respect of the ecosystem (Lairon, 2010).The strong consumer demand led to the fast increase of the number of farmers committed to this farming type. Academic literature concerning organic farming is very scarce, even if some papers dealing with quality and safety were published these last years (Dangour et al., 2010; Lairon, 2010). Moreover, there are no published results on the agricultural productions from organic farming, in comparison with products from conventional farming. In other words, similar sanitary standards are applied to both ag- ricultures, particularly regarding pesticide residues for example.The recent crisis of horse meat puts in evidence the necessity of better food traceability. In spite of the current European regu- lation, the administrative documents accompanying foodstuffs (organic or conventional farming) only inform about the country of expedition and the identity of the exporter while the country or region of production or the origins of ingredients remain unknown.

http://dx.doi.org/10.1016/j.foodcont.2014.03.0350956-7135/ 2014 Elsevier Ltd. All rights reserved.

C. Bigot et al. / Food Control 48 (2015) 123e129125

According to the International Organization for Standardization (ISO 9000-2000), traceability is a risk management tool which al- lows tracing the progress of foodstuffs (from the farm to the fork). Traceability became a constant and compulsory concern for all actors of the food chain: producers, transformers and distributors have to identify and solve critical points, realize self-monitoring, but also, inform consumers about the nature of food products (UE regulation 178/2002). It permits a quicker crisis management and a fast removal of potentially dangerous food from the market. Traceability is one of the main tools that ensure both the effective responsibility of foodstuffs manufacturers, farmers to industry of the food sector and the quality of the end product and also to improve risk estimation and manage effectiveness (Raspor, 2004). However, there is, at the moment, no real analytical tool for food traceability allowing authentication of the product origin or the farming type in a simple, fast and inexpensive way.Currently, various modern analytical techniques allow deter- mining the origin of food with a certain precision (Peres, Barlet, Loiseau, & Montet, 2007). These methods can be classied in two categories:

i) physico-chemical techniques such as Magnetic Nuclear Reso- nance (MNR), Near Infra-Red Spectroscopy (NIRS) or Stable Isotope Ratio Analysis (SIRA) and ii) biological techniques (biochemical or molecular biology involving DNA, RNA, proteins or biological molecules analyses).

The skin of fresh foods (vegetables, fruits) is not sterile and carries microorganisms or their fragments. The presence of various microorganisms depends on the external environment of the food matrices (soil ecology, spoilage, insects, diseases), but also micro- organisms brought by human activity (Sodeko, Izuagbe, & Ukhun,1987). Previous works showed that there is a link between the geographical origin of food and the structure of the food microbial ora such as wild and farmed shes (Doan, Ngoc, Dijoux, Loiseau, & Montet, 2008; Tatsadjieu et al., 2010), fruits (El Sheikha et al., 2009; El Sheikha, Bouvet, & Montet, 2011) and, more recently, on marine salts (Dufoss, Donadio, Valla, Meile, & Montet, 2013). This was performed using a molecular biology method based on the extraction, PCR amplication and DGGE separation of microbial DNA (PCR-DGGE).The main objective of the present study is to discuss the pos- sibility to apply, for the rst time, a molecular microbial ecology approach to discriminate organic from conventional food using rDNA ngerprinting of microorganisms. This work is the rst step towards the creation of an analytical tool that will allow the discrimination between foods according their farming type.

2. Materials and methods

2.1. Sampling

Mature samples of nectarines and peaches were collected from orchard and trays stored at 1 C (for fteen days for sustainable fruits, one day for organic fruits and nine days for conventional fruits) in the French cooperative Saveurs des Clos (Ille sur Tet in the south of France), during august 2011, in an aseptic way with gloves and transported into sterile bags to Cirad laboratories in Montpellier (France). Some fruits were harvested and others were sampled to the cooperative on platters. All these fruits were from the same geographical origin (nearby plots) and had the same variety: Yellow peaches are of the Corundum variety, yellow nectarines belong to the Amber variety, except conventional nectarines from platters that are of the Western Red variety. White nectarines and peaches are of Star Pearl and Amanda varieties, respectively. The fruit sampled

were of three different farming types: organic (certied by compe- tent national authority), conventional and sustainable farming (organized by the cooperative as more reasoned than the conven- tional farming). For each farming type, at least four fruits were sampled and analyzed to ensure the reproducibility of the method.

2.2. Extraction of total DNA

DNA was extracted immediately from fresh fruits skins. Total DNA was extracted by using a method adapted from Masoud, Cesar, Jespersen, and Jakobsen (2004), Ros-Chumillas, Egea-Cortines, Lopez-Gomez, and Weiss (2007) and El Sheikha et al. (2009). About2 g of peeled skin were separately mixed with 10 mL of sterile peptone water in 50 mL Falcon tubes and incubated on rotating wheel for 30 min at room temperature. Then, 1 mL of the resulting suspension was sampled in Eppendorf tubes containing about 0.3 g of acid washed glass beads (SigmaeAldrich). The mixture was vortexed vigorously for 15 min in a bead beater instrument (Vortex Genie 2) then centrifuged at 12 000 g for 15 min and the super- natant discarded. The cell pellet was resuspended in 300 mL breaking buffer [2% Triton X-100 (Prolabo), 1% SDS (sodium dodecyl sulfate; Sigma), 100 mM NaCl [(Sigma), 10 mM Tris, pH 8.0, 1 mM EDTA, pH 8.0 (Promega)]. 100 mL of buffer TE (10 mM TriseHCl pH 8,1 mM EDTA; Promega), 100 mL of lysozyme solution (25 mg/mL,Eurobio) and 100 mL of proteinase K solution (20 mg/mL, Biosolve) were successively added followed by 20 of incubation at 42 C. Then 50 mL of 20% SDS were added to each tube, and incubated at42 C for 10 min. 400 mL of 2% CTAB (cetyltrimethylammoniumbromid, Merck) were added to each tube and incubated at 65 C for10 min. The lysates were subjected twice to phenol chloroform extraction by adding 700 mL of phenol/chloroform/isoamyl alcohol mixture (25/24/1, Carlo Erba), manually mixed and then centri- fuged at 12 000 g for 15 min. The aqueous layer was transferred to a new Eppendorf tube. The residual phenol was removed by adding600 mL of chloroform/isoamyl alcohol (25:24:1, Carlo Erba) andcentrifuged at 12 000 g for 10 min. The aqueous phase was collected and 0.1 volume of sodium acetate was added (3 M, pH 5), followed by addition of an equal volume of isopropanol and stored at 20 C for 12 h (overnight). After centrifugation at 12 000 g for 30 min, the supernatant was eliminated, DNA pellets were washed with 500 mL of 70% ethanol, and tubes were centrifuged at 12 000 g for 5 min. The ethanol was then discarded and the pellets were air dried at room temperature for several hours. Finally, the DNA was resus- pended in 100 mL of ultra-pure water and stored at 20 C until analysis. DNA quantities were estimated by electrophoretic migration through a 0.8% agarose gel and by using a UV spectro- photometer (BioSpec-Nano, Shimadzu). Gels were photographed on a UV transilluminator with a CCD camera and Gel Smart 7.3 system (Clara Vision).

2.3. PCR-Denaturing Gradient Gel Electrophoresis (DGGE) analysis

For yeast DNA, a fragment of the D1/D2 region of the 26S rDNA gene was amplied using eukaryotic universal primers: forward, NL1GC (50 -CGC CCG CCG CGC GCG GCG GGC GGG GCG GGG GCC ATA TCA ATA AGC GGA GGA AAA G-30 ; Sigma); reverse, LS2 (50 -ATT CCC AAA CAA CTC GAC TC-30 ; Sigma), amplifying a 250 bp fragment (Cocolin, Bisson, & Mills, 2000; Kurtzman & Robnett, 1998). For the study of bacterial community, a fragment of the V3 variable region16S DNA gene was amplied using universal bacterial primers: forward, gc338F (50 CGC CCG CCG CGC GCG GCG GGC GGG GCG GGG GCA CGG GGG GAC TCC TAC GGG AGG CAG CAG 30 ; Sigma); reverse, 518R (50 ATT ACC GCG GCT GCT GG 30 ), amplifying a 190 bp fragment (Ampe, Ben Omar, Moizan, Wacher, & Guyot, 1999; Leesing, 2005; Le Nguyen, Hanh, Dijoux, Loiseau, & Montet,

2008). PCR amplication was performed in a nal volume of 50 mL containing 0.2 mM of each primer, all the deoxyribonucleotide triphosphate (dNTPs, Promega) at 200 mM, 5 mL of 10 of TopTaq reaction buffer containing 15 mM MgCl2 (Qiagen), 1.25 U TopTaq DNA polymerase and 5 mL of extracted DNA ( 1 mg/reaction). The amplication was carried out as follows: An initial denaturation at94 C for 3 min, 35 cycles of 94 C for 30 s, 52 C (for yeast) or 55 C (for bacteria) for 30 s, 72 C for 1 min and a nal extension at 72 C for 10 min. Aliquots (5 mL) of PCR products were analyzed rst by electrophoresis in 2% (w/v) agarose gel with TAE 1 buffer (40 mM TriseHCl, pH 7.4, 20 mM sodium acetate, 1.0 mM Na2-EDTA), stained with ethidium bromide in TAE 1 for 10 min and rinsed with distilled water for 5 min. Gels were photographed and PCR bands quantied by using a standard DNA (100 bp ladder, Prom- ega). PCR products were then separated by Denaturing Gradient Gel Electrophoresis (DGGE) [(Dcode TM universal mutation detection system, Bio-Rad)], using the procedure rst described by Muyzer, De Waal, and Uitterlinden (1993) and improved by Leesing (2005). Similar amounts of PCR amplicons were loaded into 8%(w/v) polyacrylamide gels (acrylamide/N,N-methylene bisacryla- mide, 37.5/1, Promega) in 1 TAE buffer (40 mM TriseHCl, pH 7.4,20 mM sodium acetate, 1.0 mM Na2-EDTA). Electrophoreses were performed at 60 C, using a denaturing gradient in the 30e60% range (100% corresponding to 7 M urea and 40% v/v formamide, Promega). The gels were electrophoresed at 20 V for 10 min and then at 80 V for 12 h. After electrophoresis, the gels were stained for30 min with ethidium bromide, rinsed in distilled water for 20 min and then photographed as described above.Reference DNAs (Pichia sorbitophila and Candida apicola for yeast, Escherichia coli and Lactobacillus plantarum for bacteria) were used to record DNA band position in the patterns obtained.

2.4. Image and statistical analysis

Individual lanes of gel images were aligned and processed using ImageQuant TL software version 2007 (Amersham Biosciences). This software allows detection, precise measure and record of the relative position of each DNA band. The DGGE banding pattern is considered as an image of all of the major microbial species in the population, yeast and bacterial species in this case. An individual discrete band refers to a unique sequence type or phylotype (Kowalchuk et al., 1997; Muyzer, Hottentrger, Teske, & Wawer,1996; Van Hannen et al., 1999), which is treated as a discrete yeast or bacteria population.

2.4.1. Dice similarity coefcientDGGE ngerprints were manually scored by presence and absence of co-migrating bands between lanes, independent of in- tensity. Pairwise community similarities were quantied using the Dice similarity coefcient (SD) (Heyndrickx, Vauterin, Vandamme, Kersters, & De Vos, 1996):

SD 2Nc =Na Nb

where Na represents the number of bands detected in sample A, Nb the number of bands in sample B, and Nc the number of bands common to both samples. The similarity index was expressed within a range of 0 (completely dissimilar) to 100 (perfect simi- larity). Dendograms were constructed using the Statistica version 6 software (StatSoft). Signicant differences of microbial commu- nities of fruits were determined by factorial correspondence anal- ysis, using the rst two factors that described most of the variation in the data set.A Cluster Analysis was performed using the similarity matrix to group samples according to their similarity index. The

reconstruction method used was group average by using Primer v.6 software (Primer-E Ltd).

2.4.2. Partial least squares discriminant analysis (PLS-DA)PLS-DA is a regression technique which maximizes the separa- tion between pre-dened classes. The aim is to predict the values of a group of variables X (dependent variables) from a set of variables Y (explanatory variables). In our case, Y represents quantitative variables with band volume (surface of the band in pixel multi- plied by the intensity level of each pixel of the band value measured by ImageQuant TL software) and X represents qualitative variables, that is to say the type of farming (organic, conventional or sus- tainable). In our study, binary classication models were developed (for example 1 for organic and 0 for conventional) and the belonging to one of the classes was predicted by PLS-DA according bacterial band intensity value. To optimize the number of PLS-DA latent components (LV), the percentage of correct classications (sensitivity and specicity) obtained was validated.The data were manipulated using MATLAB software Version 7.6 (R2008a).So, PLS-DA allows to combine variables in the data set to nd the maximum correlation between them and the class variable and, thus, the maximum separation among classes (organic vs. con- ventional vs. sustainable).

3. Results and discussion

Yeast and bacterial ecologies were analyzed on nectarines and peaches from various farming types in order to estimate the impact of the agricultural practices on the microbial ecology structure. In order to minimize the effect of parameters that are known to affect the associated microbial ora, fruits of the same variety and geographical origin were sampled in this study (El Sheikha et al.,2009, 2011).Yeast and bacterial DNA abundance on fruits were evaluated by quantitative PCR. The results obtained did not allow any signicant difference between fruit types to be determined (data not shown). This suggests that variations observed between yeast and bacterial ecologies (DGGE patterns) would not be due to differences in mi- crobial DNA abundance associated to the fruits but rather to different farming types.The comparison of microbial ecology of nectarines and peaches was realized by PCR-DGGE. For each type of fruit, the microbial ecology was analyzed on four fruits and produced identical DGGE patterns indicating a very high reproducibility (see samples 1e4 for each farming type in Figs. 1 and 2). Co-migrating bands as well as overall pattern variations between samples were clearly visible and further evaluated by Principal Component Analysis (PCA). Nectarines and peaches replicates showed identical yeast and bacterial rDNA DGGE patterns. Remarkably, the analysis of yeast ecology showed variations between samples according to their farming type (Figs. 3 and 4), but also according to their level of freshness (cooperative vs. orchard, Fig. 4.2). Particularly, there is a signicant differentiation in between conventional fruits compared to the fruit types (Fig. 4.1 and 4.2), with a pool of con- ventional nectarines and conventional peaches for 90% or 80% of similarity respectively.Thus, PCA allowed to group samples according to their farming type by the analysis of yeast ecology of nectarines and peaches.The analysis of bacterial ecology of fruits showed variations that could not lead to discrimination using PCA (data not shown) despite the visible signicant differences between bacterial di- versity of organic and conventional fruits (for nectarines, either white or yellow; see arrows in Fig. 2). Another statistical method was applied taking into account the variations in band intensities

Fig. 1. PCR-DGGE band patterns of yeast 26S rDNA of nectarines and peaches samples of three different farming types: organic, sustainable and conventional.

that were observed in the different DGGE proles between fruit types. When Dice similarity coefcient is not enough discrimi- nating, partial least squares discriminant analysis (PLS-DA) could exploit the information obtained from DGGE data and to reveal subtle differences that might exist between different types of samples. PLS-DA is commonly used in chemometrics for spectro- metric analysis but not often use in the area of genetics. It was notably applied in microarray data (Datta, 2001; Nguyen & Rocke,2002) or Near Infra-Red Spectroscopy data (Tres, Van Der Veer, Perez-Marin, Van Ruth, & Garrido-Varo, 2012), but not using ngerprint obtained after PCR-DGGE. When PCA is used as a multivariate technique for exploratory data analysis, PLS-DA is a discriminant method that allows nding the maximum separation among pre-dened classes (farming types in our case).

This methodology was applied to analyze data obtained taking account the intensity of each bands present on the DGGE gels for nectarines and peaches (Figs. 5 and 6).Applying this methodology to the different classes of nectarines (two classes: organic vs. conventional) and peaches (four classes: organic vs. conventional vs. platters vs. orchards) we found a good class prediction (Fig. 5): a good separation of the two types of nectarines was obtained (Fig. 5.1) and especially, it was found that conventional fruits were most distinguishable from other types (Fig. 5.2), as previously mentioned for the analysis of yeast ecology of peaches.The graphical representation of scores, for nectarines (Fig. 6.1) and peaches (Fig. 6.2), on the rst three PLS components conrmed these results: the graph of scores of the rst two latent

Fig. 2. PCR-DGGE band patterns of bacterial 16S rDNA of nectarines and peaches samples of three different farming types: organic, sustainable and conventional: (/) Bacillus.

Fig. 3. Factorial variance analysis of yeast (26S rDNA) DGGE patterns of nectarines (1) and peaches (2) samples of three different farming types.

Fig. 4. Cluster analysis of yeast 26S rDNA DGGE patterns of nectarines (1) and peaches (1) samples from three different farming types (Conv. conventional).

variables in Fig. 6.1 shows that it was possible to observe the formation of two groups, conventional and organic fruits. The graph of scores of the rst three latent variables (Fig. 6.2) showed that sustainable fruits seems to be nearest to organic than con- ventional fruits as it was previously observed for the analysis of yeast ecology (Fig. 3.1). Furthermore, the inuence of the level of freshness (cooperative vs. orchard) was also observed for the bacterial ecology (Fig. 6.2).So, the bacterial ecology analysis was also useful for the discrimination of nectarines and peaches according their farming

type. Because subtle differences were not detected by using PCA the application of PLS-DA was necessary to exploit the information from DGGE analyses.The analysis of variable importance in projection (VIP) was also performed and showed the most important variables considered for the model (data not shown). The VIP scores can be employed as a criterion for the selection of organic markers. The trends of VIPs allow the identication of the most discriminating variables, among which the two bands observed on organic nectarines and peaches DGGE gels (see arrows in Fig. 2).

Fig. 5. Prediction in PLS-DA model of nectarines (1) and peaches samples (2) to discriminate the farming type by the analysis of the bacterial ecology.

Fig. 6. Plot of the scores of LV1 LV2 for nectarines (1) and of LV1 LV2 LV3 for peaches (2) to discriminate the farming type by the analysis of the bacterial ecology.

The sequencing analysis of these bands (represented by arrows) showed that the corresponding bacteria belong to the Bacillus ce- reus group (including Bacillus weihenstephanensis and Bacillus thuringiensis). Remarkably, these species are routinely used in biological pest control like in organic farming. In order to test this bacterial species as potential markers that could be used to discriminate organic from other types of foods, it could be judicious to apply quantitative PCR with specic primers previously designed according sequences of species found in this study.Thus, the PLS-DA has proven to be effective because it gives us an alternative to be able to analyze subtle differences and to identify the most discriminative bands, and so the most discrimi- native bacterial species detected by DGGE analysis.In all, we showed that the analysis of the microbial ecology of peaches and nectarines allowed them to be differentiated accord- ing their farming type. The differences must be assessed with suitable statistical tools to identify discriminant markers to allow the creation of detection tools for control, authentication and traceability.

4. Conclusions

The main objective of this work was to test the possibility of discrimination of food production modes by the study of their microbial ecology. It was based on the hypothesis that treatments associated to various farming types have a measurable effect on food microora.Most studies dealing with discrimination of organic from con- ventional foods use isotopic, chemical and biochemical analysis. The present study showed that the application of molecular mi- crobial ecology approaches (such as PCR-DGGE) could serve as discrimination tools using bacterial and yeast rDNA markers in a quick and cost effective way. Indeed, the analysis of yeast and bacterial genetic proles of nectarines and peaches showed that it was possible to differentiate fruits from different farming types. The differences observed in the microbial ecology were accurate enough to conclude that they resulted exclusively from applied treatments and from the level of freshness.It is possible that, during post-harvest stages, the microbial DNA ngerprint changes (e.g. due to food storage and transport). The aim of the present study was to test the possibility of differentiating organic fruits from conventional fruits when these were harvested from orchards or from platters and signicant differences were observed. For further studies, and the creation of molecular tools for authentication and control, it could be interesting to verify the robustness of bacterial and yeast proles (or biological barcode)

along seasons or years to be used as reliable traceability markers for organic and/or non-organic foods.

Acknowledgements

The authors would like to thank Mr MALATERRE Dominique and Mr FRISSANT Rmy (France) for the supply of nectarines and peaches samples.Celine Bigot is supported by a PhD grant as well as a DESI grant(2000V) from CIRAD.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.foodcont.2014.03.035.

References

Ampe, F., Ben Omar, N., Moizan, C., Wacher, C., & Guyot, J. P. (1999). Polyphasic study of the spatial distribution of microorganisms in Mexican pozol, a fermented maize dough, demonstrated the need for cultivation-independent methods to investigate traditional fermentations. Applied and Environmental Microbiology,65, 5464e5473.Cocolin, L., Bisson, L. F., & Mills, D. A. (2000). Direct proling of the yeast dynamics in wine fermentations. FEMS Microbiology Letters, 189, 81e87.Dangour, A. D., Lock, K., Hayter, A., Aikenhead, A., Allen, E., & Uauy, R. (2010).Nutrition-related health effects of organic foods: a systematic review. AmericanJournal of Clinical Nutrition, 92(1), 203e210.Datta, S. (2001). Exploring relationships in gene expressions: a partial least squares approach. Gene Expression, 9, 249e255.Doan, L. N., Ngoc, H. H., Dijoux, D., Loiseau, G., & Montet, D. (2008). Determination of sh origin by using 16SrDNA ngerprinting of bacterial communities by PCR- DGGE: an application on Pangasius sh from VietNam. Food Control, 19, 454e460.Dufoss, L., Donadio, C., Valla, A., Meile, J. C., & Montet, D. (2013). Determination of speciality food salt origin by using 16S rDNA ngerprinting of bacterial com- munities by PCR-DGGE: an application on marine salts produced in solar salt- erns from the French Atlantic Ocean. Food Control, 32, 644e649.El Sheikha, A. F., Bouvet, J. M., & Montet, D. (2011). Biological bar code for deter- mining the geographical origin of fruits using 28S rDNA ngerprinting of fungal communities by PCR-DGGE: an application to Shea tree fruits. Quality Assurance and Safety of Crops & Foods, 3, 40e47.El Sheikha, A. F., Condur, A., Mtayer, I., Le Nguyen, D. D., Loiseau, G., & Montet, D. (2009). Determination of fruit origin by using 26S rDNA ngerprinting of yeast communities by PCR-DGGE: preliminary application to Physalis fruits from Egypt. Yeast, 26, 567e573.Heyndrickx, M., Vauterin, L., Vandamme, P., Kersters, K., & De Vos, P. (1996).Applicability of combined amplied ribosomal DNA restriction analysis (ARDRA) patterns in bacterial phylogeny and taxonomy. Journal of Microbio- logical Methods, 26, 247e259.Kowalchuk, G. A., Stephen, J. R., De Boer, W., Prosser, J. I., Embley, T. M., & Woldendorp, J. W. (1997). Analysis of ammonia-oxidizing bacteria of the beta subdivision of the class proteobacteria in coastal sand dunes by denaturing gradient gel electrophoresis and sequencing of PCR-amplied 16S ribosomal DNA fragments. Applied and Environmental Microbiology, 63, 1489e1497.

Kurtzman, C. P., & Robnett, C. J. (1998). Identication and phylogeny of ascomyce- tous yeasts from analysis of nuclear large subunit (26S) ribosomal DNA partial sequences. Antonie Van Leeuwenhoek, 73, 331e371.Lairon, D. (2010). Nutritional quality and safety of organic food. A review. Agronomy for Sustainable Development, 30, 33e41.Leesing, R. (2005). Identication and validation of specic markers for traceability of aquaculture sh for import/export (Ph.D. dissertation). France: University of Montpellier 2.Le Nguyen, D. D., Hanh, H. N., Dijoux, D., Loiseau, G., & Montet, D. (2008). Deter- mination of sh origin by using 16S rDNA ngerprinting of bacterial commu- nities by PCR-DGGE: an application on Pangasius sh from Viet Nam. Food Control, 19, 454e460.Masoud, W., Cesar, L. B., Jespersen, L., & Jakobsen, M. (2004). Yeast involved in the fermentation of Coffea arabica in East Africa determined by genotyping and by direct denaturing gradient gel electrophoresis. Yeast, 21, 549e556.Muyzer, G., De Waal, E. C., & Uitterlinden, A. G. (1993). Proling of complex mi- crobial populations by denaturing gradient gel electrophoresis analysis of po- lymerase chain reaction-amplied genes coding for 16S rRNA. Applied and Environmental Microbiology, 59, 695e700.Muyzer, G., Hottentrger, S., Teske, A., & Wawer, C. (1996). Denaturing gradient gel electrophoresis of PCR-amplied 16S rDNA e a new molecular approach to analyse the genetic diversity of mixed microbial communities. Molecular Mi- crobial Ecology Manual, 34, 1e23.Nguyen, D. V., & Rocke, D. M. (2002). Tumor classication by partial least squares using microarray gene expression data. Bioinformatics, 18, 39e50.

Peres, B., Barlet, N., Loiseau, G., & Montet, D. (2007). Review of the current methods of analytical traceability allowing determination of the origin of foodstuffs. Food Control, 18, 228e235.Raspor, P. (2004). Bio-markers as primary identiers as needed for food safety and traceability of food items. In 2nd Central European congress on food (pp. 1e11). CEFood congress. Consumers, nutrition, safety, technology.Ros-Chumillas, M., Egea-Cortines, M., Lopez-Gomez, A., & Weiss, J. (2007). Evalua- tion of a rapid DNA extraction method to detect yeast cells by PCR in orange juice. Food Control, 18, 33e39.Sodeko, O. O., Izuagbe, Y. S., & Ukhun, M. E. (1987). Effect of different preservative treatment on the microbial population of Nigerian orange juice. Microbios, 51,133e143.Tatsadjieu, N. L., Mawor, J., Hadjia, M. B., Loiseau, G., Montet, D., & Mbofung, C. M. F. (2010). Study of the microbial diversity of Oreochromis nilo- ticus of three lakes of Cameroon by PCR-DGGE: application to the determination of the geographical origin. Food Control, 21, 673e678.Tres, A., Van Der Veer, G., Perez-Marin, M. D., Van Ruth, S. M., & Garrido-Varo, A. (2012). Authentication of organic feed by near-infrared spectroscopy combined with chemometrics: a feasibility study. Journal of Agricultural and Food Chem- istry, 60, 8129e8133.Van Hannen, E. J., Zwart, G., van Agterveld, M. P., Gons, H. J., Ebert, J., & Laanbroek, H. J. (1999). Changes in bacterial and eukaryotic community struc- ture after mass lysis of lamentous cyanobacteria associated with viruses. Applied and Environmental Microbiology, 65, 795e801.