3. Authentication of fishery products

Species authentication is a key element for verifying foodstuff fairness and fighting deceptive practices. This is true in particular for fishery products due to the huge number of species which can be used as raw material and to the great diversity of preparations marketed.

In section 11.2 “Responsible international trade”, the FAO Code of Conduct for Responsible Fisheries (http://www.fao.org/fi/agreem/codecond/codecon.asp) calls for the liberalization of trade in fish and fishery products and for the elimination of unjustified barriers, in accordance with the principles laid down in the agreements of the World Trade Organization. But such liberalization can only take place in a framework of transparency and enhanced information to consumers, particularly with regard to product labelling.

The Codex standards have become an integral part of the legal framework within which international trade is being facilitated through harmonization. They have already been used as the benchmark in international trade disputes, and it is expected that they will be used increasingly in this regard. The adoption of Codex standards as scientifically justified norms for the purpose of the World Trade Organization Sanitary and Phytosanitary (SPS) and Technical Barrier to Trade (TBT) Agreements is of immense significance. Fish exporting countries are increasingly seeking recognition of their fishery products in the Codex standards. The potential reward from including additional species or families of species in a Codex standard is of course linked to the international recognition of the product in question, and it is perfectly legitimate for a country to want to derive maximum benefit from its resources and expertise. This recognition is primarily associated with the commercial name of the product; authorization to use a name with established international repute is therefore an important asset and a declared objective. However, there are many species seeking value-enhancing appellations, but such appellations are relatively few. Labelling provisions therefore need to be sufficiently clear to avoid misleading consumers and creating conditions of unfair competition in international trade.

The finalization or amendment of fishery product standards that are credible and authoritative requires extensive consultation as well as the collection and evaluation of information, followed up by confirmation of final results and sometimes objective compromise to satisfy differing sound, scientifically based views. With the prospect of growing international trade and an increasing number of potentially marketable species, it is worth considering some legislative issues in some major markets and conducting a review of the modern analytical techniques to verify product authenticity.

In the following two examples improper labelling might have been more damaging to the producer than to the consumer, because the products may acquire a reputation as second class products. It is easier to persuade a person to try a new food product than to convince them to change their opinion about the quality of a product that had been introduced under fraudulent premises in the local market.

People in the Basque region have a high regard for their own fisheries, which include “legatza” (merluza, hake, Merluccius merluccius). The catches of this species have decreased, and its price increased. This has encouraged the introduction into the market of other hake species including Merluccius australis, M. capensis, M. gayi, M. hubbsi, M. polylepis and Macruronus novaezelandie. This introduction, unfortunately, has sometimes been fraudulent: purchasers were led to believe that they were buying the “real” hake. The first impression when the fraud was discovered was that they were being fooled and given a lower quality product, which is not necessarily true: in the opinion of a fisherman who had tasted the “alternative” hake when fishing it, the organoleptic properties of the “new” were not at all inferior to the traditional one. A second problem is that a whole series of different species were introduced under only one name, which also prevented customers from getting to know and appreciate them individually. Although the situation regarding correct information has improved, customers still have the impression that the “substitute” is not as good as the “real” product. If these new species had been introduced with their own names and characteristics from the beginning, customers would have become used to them as new products, and the prices of the different commodities would adjust themselves according to preferences and supply.

The second example concerns some capsules containing fish oils, marketed as containing Norwegian salmon oils, and shown to contain oils from pelagic fish from the South American Pacific coast (Aursand et al., 1999). From the point of view of quality and use as a nutraceutical, the marine oil in the fraudulent product was not of inferior composition or quality than in the original product, but the right of the customers to correct information had been breached and they would automatically assume that the mislabelled product was of inferior quality.

In both cases it is not only the purchaser who suffers damage but also the primary producer. Fishermen and breeders in the regions where the “substitutes” originated (South America and Africa in the above examples) could see their incomes increased if their products had a specific market niche and a specific recognizable name. The use of their raw materials as substitutes automatically and unnecessarily diminishes their perceived value and gives them a reputation as second class products that may be very difficult to overcome. In all these cases, the only one to obtain a profit is the middleman who knows what he purchases and then mislabels it.

3.1 Traceability and authentication

Traceability documentation is the paper trail that must accompany a product and contain all pertinent data to trace its origin and life-history. It includes species, ingredients, origins, manufacturing processes, temperature logs, etc. Traceability is of major relevance to producers, so that if a problem arises during manufacture it is possible to recall only the batch affected, and not the entire production. It is especially important to the producer if it permits recall or stops production before the affected lot leaves the premises, thus avoiding potential damage to the reputation of the company by informing the public about the problem.

Full traceability will of course help to follow up honestly produced and handled products, but because they are honestly produced and treated it is to be expected that they will also be correctly labelled. There are many reports in newspapers on cases of fraud where labels are either replaced or covered with new ones. The products appear to be perfectly traceable, but are in fact fraudulent. This sort of procedure has been used to change the nationality of eggs and chickens, relabel out-of-date frozen salmon as fresh with a new “use by” date, change expired frozen smoked trout into fresh smoked salmon, etc.

Authentication analyses, on the other hand, are analyses performed on a sample to assess whether the product is indeed what the label says: if it says Rioja wine, then an authentication analysis will show that the content of the bottle is Rioja wine. If the bottle is fraudulently labelled, the results will show a different fingerprint corresponding to a different liquid. The identity of the liquid may be identified, or not, depending on whether the fingerprint of the liquid is included in the database of results (the official method for wines in the European Union is site-specific natural isotope fractionation by nuclear magnetic resonance (SNIF-NMR), Official Journal of the European Community (1990) 33 L-272. 3 October, 1990).

So, in a way, traceability is a tool to help producers and authorities by on the one hand diminishing the risk of a faulty batch reaching the market, and on the other, if the faulty batch does reach the market, by limiting the removal to the damaged units only in order to save time, money and foodstuff.d units only. Authentication analyses on the other hand can be considered to be aimed at protecting the consumer’s right to correct information and to helping authorities, among other things, to allocate correct taxes and import/export permits for the goods.

3.2 Regulatory aspects

3.2.1 Legislation in Europe

The directives and regulations in Europe that define the rules for labelling of foods, including fishery products are:

• Directive 2000/13/EC of the European Parliament and of the Council of 20 March 2000 on the approximation of the laws of the Member States relating to the labelling, presentation and advertising of foodstuffs;

• EU Commission Regulation No 2065/2001 of 22 October 2001 (see also the Corrigendum published in OJ L 278 of 23.10.2001) establishing detailed rules for the application of Council Regulation (EC) No 104/2000 as regards informing consumers about fishery and aquaculture products;

• Council Regulation No 2081/92 of 14 July 1992 on the protection of geographical indications and designations of origin for agricultural products and foodstuffs (introducing Protected Designation of Origin (PDO), Protected Geographical Indication (PGI), Organic or Ecological, and “Traditional Speciality Guaranteed” or Specific Character labels);

• Council common position for adopting a European Parliament and Council regulation on traceability and labelling of genetically modified organisms and the traceability of food and feed products produced from genetically modified organisms and amending Directive 2001/18/EC (A5-0204/2003; 15798/1/2002 -C5-0131/2003 - 2001/0180(COD)). The final version of this text is expected to become consolidated in 2003.

Directive 2000/13/EC of the European Parliament and of the Council of 20 March 2000 on the approximation of the laws of the Member States relating to the labelling, presentation and advertising of foodstuffs includes the following premises:

In the second article, it is specified that the labelling must not be misleading to the purchaser, among other properties, as to: “the characteristics of the foodstuff and, in particular, to its nature, identity, properties, composition, quantity, durability, origin or provenance, method of manufacture or production”. Article number 3 gives a list of particulars compulsory on the labelling of foodstuffs, including (number 2) “list of ingredients”. Number 8 requires “particulars of the place of origin or provenance, where failure to give such particulars might mislead the consumer to a material degree as to the true origin or provenance of the foodstuff”.

TABLE 7

Catch area and identification of the area*

* Commission Regulation (EC) No 2065/2001 of 22 October 2001 Laying down detailed rules for the application of Council Regulation (EC) No 104/2000 as regards informing consumers about fishery and aquaculture products. Published in the Official Journal of the European Communities. Date: 23.10.2001. Pages: L278/6 to L278/8)

(¹) FAO yearbook. Fishery Statistics. Catches. Vol. 86/1. 2000

(²) Excluding the Baltic Sea

However, the ingredient may be designated by the name of the category, i.e., “fish”, rather than by the specific name for “all species of fish where the fish constitutes an ingredient of another foodstuff and provided that the name and presentation of such foodstuff does not refer to a specific species for fish”.

European consumers are also entitled to additional information when purchasing fish products: thus, the EU Commission regulation No 2065/2001 of 22 October 2001(see also the Corrigendum published in OJ L 278 of 23.10.2001) has established detailed rules for the application of Council Regulation (EC) No 104/2000 with regard to informing consumers about fishery and aquaculture products. The information includes specification of the commercial designation and scientific name, method of production of a species (“caught” or “caught in freshwater” or “farmed” or “cultivated”) and the area in which it was caught (Table 7). In the case of cultivated species, article 5 of the regulation indicates that a reference should be made to the country in which the product undergoes the final developmental stage. Since 2000 some European countries have demanded that similar labelling on seafood be available to consumers.

Protected (PDO, PGI) and traditional designations in Europe

The European Union has established legislation for the protection of geographical indications and designations of origin for agricultural products and foodstuffs in order to encourage a greater diversity of agricultural production; protect product names from misuse and imitation, and give the necessary information to consumers about the specific character and origins of products (Council Regulation (EEC) No 2081/92 of 14 July 1992) by introducing Protected Designation of Origin (PDO), Protected Geographical Indication (PGI), Organic or Ecological, and “Traditional Speciality Guaranteed” or Specific Character labels). These denominations were first introduced at the national level to protect agricultural products or foodstuffs which have an identifiable geographical origin and have proved successful with producers, who have secured higher incomes in return for a genuine effort to improve quality, as well as with consumers, who can purchase high quality products with guarantees as to the method of production and origin.

Also Norway, not a EU member and with little tradition for this kind of labelling, has recently introduced similar legislation for agricultural products (FOR 2002-07-05 nr 698) and it can be reasonably assumed that the regulation will soon extend to fish and fishery products (Ot. prp. nr. 39 [2002-2003]). In Europe the legislation already covers fishery products.

The EU’s agriculture ministers have updated the regulations designed to ensure the authentic use of geographical sources and origin for food production (EC-DGA, 2003). The update has three important objectives:

• allow producers to include in their product specifications the requirement that packaging of products must take place in a defined geographical area;

• bring the EU’s regulation in line with the WTO’s intellectual property rules and make it possible for third countries to apply for products to be registered under the EU system; and

• include new products such as mustard and pasta in the scope of the regulation.

The United States and other countries have challenged the EU’s PGI rules in the World Trade Organization. The European Commission is mounting a vigorous defence, because it is considered that improving food quality is central to the prospects of farmers and also to protect consumers’ rights to correct information, as they seek products with special quality characteristics such as geographical specified origin or particular production methods. The Commission does not want consumers to be misled or that products named after their true origin be excluded from the market because geographical designations may be used as trademarks. For example, Parma ham could be registered as a trademark in, say Canada, and then real Parma ham could no longer be sold under its real name (taken from EC-DGA, 2003).

Labelling of GM products

Another type of compulsory label in the EU refers to food and feed products produced from genetically modified organisms (GMOs) and to the traceability of such products (Council common position for adopting a European Parliament and Council regulation on traceability and labelling of genetically modified organisms and the traceability of food and feed products produced from genetically modified organisms and amending Directive 2001/18/EC (A5-0204/2003; 15798/1/2002-C5-0131/2003-2001/0180(COD)).

According to the preliminary text, the European Council has set a limit of 0.9 percent, above which food containing GMO material whose presence is adventitious or technically unavoidable would have to be labelled accordingly. If live GMOs are present in products they should be labelled irrespective of whether their presence is adventitious or technically unavoidable. These thresholds refer to authorized GMOs; there is zero tolerance for unauthorized GMOs. In order to ensure that the market will function correctly and that customers will receive accurate information, the Council attempts to focus mainly on the traceability of foods and feeds, particularly for those cases where it is not possible to use analytical methods to distinguish between conventional products and those produced from GMOs. One would expect that improved analytical methods and more stringent production processes will lead to lower tolerance values and to more accurate determination of contamination values. Currently, it is possible to detect and quantify gene modified material in foods (Wurz et al., 1999; see also a recent review by Ahmed, 2002). Kuribara et al. (2002) have applied the technique to the detection of gene modifications in maize and soybean, and Brodmann et al. (2002) to four different varieties of maize approved in the European Union: insect-resistant Bt 176 maize (Maximizer), Btl 1 maize, Mon 810 (Yield Gard) maize, and the herbicide-tolerant T25 (Liberty Link) maize. Ankilam et al. (2002) have published the results of validation studies for the detection and quantification of GM material in foods. Their work shows the need for further validation studies, in particular regarding the application of quantitative PCR methods using new extraction methods, diverse matrixes, and novel GMO constructs.

Although there are yet no commercially available gene modified fish in the market, there are already several stable lines of genetically modified fish: some with additional genes to make the fish tolerant to cold temperatures, or to increase its growth rate (duplicating the fish’s own growth hormone gene), etc. It is reasonable to assume that such fish will soon be on the market in countries open to these products.

Not only genetically modified fish, but also fish fed on genetically modified products will have to be labelled in the European community, since the new legislation will make compulsory the traceability and the labelling of GMOs in feeds, as well as foods. Thus, fish fed on diets containing GM components, for example genetically modified soya or maize, will also have to be labelled.

A great majority of European customers will want information regarding species and harvesting area and production methods in their foodstuffs. Factors of relevance in deciding whether to purchase a product and also to estimate a fair price include the fishing gear used for capture, fishing grounds, landing harbour, freshness, if the product is produced following environmentally friendly, ecological or traditional methods, if gene technology has been used, if wild or cultivated, etc.

3.2.2 Legislation in the United States of America

The Food and Drug Administration (FDA) is responsible for assuring that foods sold in the United States are safe, wholesome and properly labelled. This applies to foods produced domestically, as well as foods from foreign countries. The Federal Food, Drug, and Cosmetic Act and The Fair Packaging and Labeling Act are the Federal laws governing food products under FDA’s jurisdiction”.

The Fair Packaging and Labeling Act, Title 15 - Commerce and Trade, Chapter 39 - Fair Packaging and Labeling Program §145; Congressional Delegation of Policy, establishes that: “Informed consumers are essential to the fair and efficient functioning of a free market economy. Packages and their labels should enable consumers to obtain accurate information as to the quantity of the contents and should facilitate value comparisons. Therefore, it is hereby declared to be the policy of the Congress to assist consumers and manufacturers in reaching these goals in the marketing of consumer goods”.

The Labeling Act states that the commodity shall bear a label specifying the identity of the commodity and the name and place of business of the manufacturer, packer, or distributor. On the labelling of packaged food, it states that:

(a) The principal display panel of a food in packaged form shall bear, as one of its principal features, a statement of the identity of the commodity;

(b) Such statement of identity shall be in terms of: (1) The name now or hereafter specified in or required by any applicable Federal law or regulation; or, in the absence thereof: (2) The common or usual name of the food; or, in the absence thereof: (3) An appropriately descriptive term, or when the nature of the food is obvious, a fanciful name commonly used by the public for such food.

In addition, nutritional information must be provided on the label for almost all packaged food products intended for sale to the consumer. The regulations specify the nutrition information that must be on the label and the format in which it is to be presented. Regulations also prescribe conditions under which nutrient content claims or health claims may be made on the label or in labelling for a food product.

“The Seafood List” has been elaborated for the correct labelling of sea foods. The list, developed by the FDA in co-operation with the National Marine Fisheries Service (NMFS), is a compilation of existing acceptable market names for imported and domestically available seafood. The purpose of the list is to promote uniformity in the use of FDA acceptable market names by the fish industry, and to provide consistent advice on these names. The list represents an extensive, although not necessarily complete, listing of seafood commonly sold in the United States. The Seafood List can be found at http://www.cfsan.fda.gov/~frf/seaintro.html.

There does not seem to be a need to label products with the origin or production method in the United States. This implies no need to label products in whose manufacture gene technology has been used. However, The Food and Drug Administration has issued a draft document “Guidance Industry. Voluntary Labeling Indicating Whether Foods Have or Have Not Been Developed Using Bioengineering”. The document may be found at http://www.cfsan.fda.gov/~dms/biolabgu.html.

The first difference between Europe and the United States is found in the terminology used: while “genetically modified” and “GMO” and derivatives are common in Europe, they do not seem to be recognized by consumers in the United States. Quoting the guidelines mentioned above which refer to the United States: “Terms that are frequently mentioned in discussions about labelling foods with respect to bioengineering include “GMO free” and “GM free”. “GMO” is an acronym for “genetically modified organism” and “GM” means “genetically modified”. Consumer focus group data indicate that consumers do not understand the acronym “GMO” and “GM” and prefer label statements with spelled out words that mean bioengineering (Levy and Derby, 2000).

The FDA has required special labelling on the basis of it being “material” information in cases where the absence of such information may: 1) pose special health or environmental risks (e.g. warning statements on protein products used in very low calorie diets); 2) mislead the consumer in the light of other statements made on the label (e.g. requirement for quantitative nutrient information when certain nutrient content claims are made about a product); or 3) in cases where a consumer may assume that a food, because of its similarity to another food, has nutritional, organoleptic, or functional characteristics of the food it resembles when in fact it does not (e.g. reduced fat margarine not suitable for frying). The label of the food must reveal all “material” facts about the food.

Although the 1992 policy does not require special labelling for bioengineered foods, the agency advised in that policy that labelling requirements that apply to foods in general also apply to foods produced using biotechnology. Section 403(i) of the Federal Food Drug, and Cosmetic Act requires that each food bear a common or usual name or, in the absence of such a name, an appropriately descriptive term. In addition, under section 201(n), the label of the food must reveal all material facts about the food. So, if the “new” product has a “new” property, the property must be referred to in the label, to avoid having customers confuse the old and new products. However, the way or methods used in the manufacture of the new product (genetic engineering in this case), is not compulsory information. Thus:

• if a bioengineered food is significantly different from its traditional counterpart such that the common or usual name no longer adequately describes the new food, the name must be changed to describe the difference;

• if an issue exists for the food or a constituent of the food regarding how the food is used or consequences of its use, a statement must be made on the label to describe the issue;

• if a bioengineered food has a significantly different nutritional property, its label must reflect the difference;

• if a new food includes an allergen that consumers would not expect to be present based on the name of the food, the presence of that allergen must be disclosed on the label.

In determining whether a food is misbranded, FDA would review label statements about the use of bioengineering to develop a food or its ingredients under sections 403(a) and 201(n) of the Act. Under section 403(a) of the Act, a food is misbranded if statements on its label or in its labelling are false or misleading in any particular. Under section 201(n), both the presence and the absence of information are relevant to whether labelling is misleading.

The following are examples of some statements that might be used. The discussion accompanying each example is intended to provide guidance as to how similar statements can be made without being misleading.

• Genetically engineered or This product contains cornmeal that was produced using biotechnology. The information that the food was bioengineered is optional and this kind of simple statement is not likely to be misleading. However, focus group data indicate that consumers would prefer label statements that disclose and explain the goal of the technology (why it was used or what it does for/to the food). Consumers also expressed some preference for the term “biotechnology” over such terms as “genetic modification” and “genetic engineering” (Levy and Derby, 2000).

• This product contains high oleic acid soybean oil from soybeans developed using biotechnology to decrease the amount of saturated fat. This example includes both required and optional information. As discussed above in the background section, when a food differs from its traditional counterpart such that the common or usual name no longer adequately describes the new food, the name must be changed to describe the difference. Because this soybean oil contains more oleic acid than traditional soybean oil, the term “soybean oil” no longer adequately describes the nature of the food.

• These tomatoes were genetically engineered to improve texture. In this example, the change in texture is a difference that may have to be described on the label. If the texture improvement makes a significant difference in the finished product, sections 201(n) and 403(a)(1) of the Act would require disclosure of the difference for the consumer. However, the statement must not be misleading. The phrase “to improve texture” could be misleading if the texture difference is not noticeable to the consumer.

Unlike Europe, in the United States the fact that biotechnology has been used in production is irrelevant for labelling. Only new product properties must be declared.

It seems that the major difference between the philosophies behind the labelling of food products between the EU and the USA lies in what the authorities consider that their citizens need to know to make their choice. In the EU labelling is intended to ensure the consumer’s right to know any fact they (the consumers) may deem important to exercise the right to choose and pay a fair price. The US regulations, on the other hand, seem to consider that only information that will help their consumers to choose wisely (from a nutritional/compositional point of view) must be mandatory.

3.2.3 Legislation in other countries

In Canada, the Consumer Packaging and Labelling Act establishes that: “The food label is one of the most important and direct means of communicating product information between buyers and sellers and the purpose of food labelling should serve three primary functions”: a. provide Basic Product Information: the common name, list of ingredients, net quantity, durable life dates, grade/quality, country of origin and name/address of responsible manufacturer, dealer or importer; b. Provide Health/Safety and Nutrition Information: instructions for safe storage and handling, nutritional profile such as quantity of fat, protein, carbohydrates, vitamins and minerals present per serving of the food, specific information for consumers with special dietary needs; and c. Vehicle for Food Marketing, Promotion and Competition: advertises and promotes product sale and trade via label vignettes, promotional information and label claims, i.e., “low fat”, “cholesterol-free”, “high fibre”, “product of Canada”, “natural”, “organic”, “no preservatives”, etc. Label information constitutes the primary means by which consumers differentiate between individual foods and brands to make informed purchasing choices.

In Japan, the labelling of foodstuffs is regulated by The Food Sanitation Law of 1947, established for public health protection, and the Japanese Agricultural Standard Law of 1950, for product quality and labelling. As of March 2000, the standards on the labelling of fresh foodstuffs demand that the label must carry information about the name of the product, the specific name, place and country of origin. For processed foods it must also include the name of the product, name and address of the processor, the list of ingredients and additives, date of manufacture, use by date, instructions for storage, and place or country of origin. Products derived from GMOs must be labelled (Fukutomi, 2000).

3.2.4 Misleading terms

In general, the legislation of all countries indicates that the label must not mislead consumers, but international trade and the use of similar terms for different products makes it complicated when a product from one country is introduced into another in which the niche already exists. Different species may have the same common name in different countries (or regions of the same country). On the other hand, sometimes the same species receives different names in the same language at different locations in the same country. For example, the Seafood List of the FDA lists 19 different species under a search for “hake”. Eight of them have “hake” as a market name in the USA. In Spain, however, only one would be the “real hake” Merluccius merluccius, but this particular species has “whiting” as the market name and “European hake” as the common name in the USA. In Norway “bacalao” is the name of a dish made of salted and dried (clipfish) cod (Gadus morhua) and “sardiner” and “ansjos” are both canned Spratus spratus (which is called brisling if not canned). In Spain “bacalao” is the common name of Gadus morhua, and “sardina” and “anchoa” are the common names of Sardina pilchardus and Engraulis encrasicolus, respectively.

The same kind of problem appears in defining standards and names for standards (for species and products), in the Codex Alimentarius (see the Discussion paper on the procedure for the inclusion of additional species and on labelling requirements related to the “name of the product” in Codex standards. Joint FAO/WHO Food Standards Programme Codex Committee on Fish and Fishery Products, 25^th Session. Ålesund, Norway, 3-7 June 2002).

3.2.5 “Quality” labels and protected labels

In addition to compulsory information on the label, many products also contain additional labels and marks. Some of these marks are: brand labels, collective trademarks, “certificates of origin”, and “other” quality labels.

Brands usually belong to a company (or consortium) that set their own standards. Companies that have established a reputation with their customers, and wish to keep it, have also established certain standards - usually more stringent than those demanded by legislation - and deliver a certain level of quality to which customers are accustomed (Kapferer, 1997). Companies use the brand name to distinguish themselves from others delivering similar products. A major goal is that customers should use the brand name as the reference name for the particular commodity. Some examples in the international market are Thermos, Hoover, Nylon and Xerox.

Collective trademarks exist when the same brand name is used by several independent producers who agree on common rules for producing and/or marketing the product (Menard and Valceschini, 1999). According to these authors, collective trade-brands can be divided into:

i) “Simple” collective trademarks, in which there is no need of co-ownership, but there are rules established by the holders of rights that all users of the trademark (who often pay for the use of the name) must implement; and

ii) “Certified” trademarks, which are those collectively owned, often by a group of producers, that contractually agreed on a set of requirements with a legal status, so that they can be enforced by a third party (Menard and Valceschini, 1999).

Most collective trademarks are of the second type and generally lead to the definition of a “label” that ensures a product has a set of specific characteristics. The characteristics may be agreed upon by the private partners alone or, as is more common in Europe, between the authorities, producers and consumers. A certifying organization then controls and approves producers according to how they follow the standards. The certifying organization must be a formal institution, public, private, or mixed, but always external to the parties dealing with the transaction. It is very important that the certifying organization be reliable and trustworthy, because failure to comply with the standards by just one organization may damage the reputation of the trademark and of all the other companies sharing the standards. Approval is never definitive and may be withdrawn. One example is the “Label Rouge” developed in the French agricultural sector, which guarantees an objective level of high quality, based on specific production and manufacturing rules, some of which may refer to a regional tradition (Menard and Valceschini, 1999).

As already mentioned above, one of the reasons for passing legislation defining protected labels in the EU is to protect traditional products with an international reputation from having their names registered (as trademarks for example) by external actors, which would lead customers to believe that they are purchasing an authentic original product when they are not. Without protected labels the producers of the original product may find that they can no longer sell their products under their true name, protected by a reputation acquired after many years of work, and that their market can be invaded by copies (although legally sold under the name of the original products). This would probably cause a lowering of the prices of both authentic and trademark products. Only the owners of the trademark will benefit from this situation, although the profit in this case may be short-lived, as customers will sooner or later realize that they are not purchasing authentic goods and their willingness to pay will fall correspondingly.

Other quality labels may be backed up by certain specifications of the product. For example, some products carry an “organic” label, meaning that they have been produced according to certain practices. It does not say anything about whether the product itself meets the standards for “organic”, for example its pesticide content. Other “organic” labels can mean that the product itself has fulfilled the requirements of an organic certifying body. In Norway, many products carry the mark of the association of allergic persons. A television programme showed that most customers assumed that this meant that the product had been tested and declared “safe” for allergic people. In reality the label only means that the firms allowed to use the label had made a financial contribution to the said association. The difference between smoked “Salmon from Norway” and “Norwegian salmon”, both being Atlantic salmon (Salmo salar), is that the first has been reared and smoked in Norway, the second does not need to have been reared in Norway, but it has been smoked “the Norwegian way”.

The expressed need for labels among French consumers of oysters was found to differ depending on their knowledge about the products. Older people and people from the coastal cities, who are the main consumers of oysters, rely on their own knowledge of the product. They place particular value on products of the region and do not desire labels or quality marks. Inhabitants of bigger cities pay higher prices for a certain appellation, “Marennes Oléron”, although they do not know the specific characteristics of the product. The trend among distributors paralleled that of consumers: 85 percent of inland distributors, but only 36 percent of coastal ones, would rely on a certified quality approach (Erwan and Paquotte, 1998). On the other hand, in a study by Norberg (2000) such labels did not have much value to Norwegian consumers who lacked knowledge of certification marks. However, they all showed great faith in the food control authorities and were not concerned about the safety of the products. This may be because Norway is a small and relatively closed market, where consumers have access mainly to national products. Therefore, for Norwegian consumers, the fact that the products were on the market shelves was probably sufficient guarantee of their approval - and therefore certification of a minimum level of quality - by the health authorities.

It is important to realize that no matter how strict the quality controls or how high the standards, customers will only appreciate “labels” or quality signs if they are familiar with them and know what they mean. The opposite is also true: they may become accustomed to believe that some marks imply a level of quality or safety that the product does not have.

In summary, labels and quality certificates are only correctly understood and used by the consumers if they are linked to a basic understanding of what they are and what they mean. It is also very important to keep in mind that to achieve a good and solid reputation may be a long task requiring a significant investment in marketing and information. On the other hand, only one scandal - a food poisoning or mislabelling incident - may be enough to ruin a hard earned reputation.

3.3 Analytical methods for species authentication

As already mentioned, seafood in Europe must carry information about the species, geographical origin of capture or last developmental stage, whether the fish are wild or cultivated and whether they contain genetically modified material. There are at present validated methods for species identification and for detection of some genetically modified material. There are, however, no validated protocols for full authentication, for example to demonstrate where the fish comes from or the processing to which it has been submitted, or perhaps if the label claims that the product has been organically (ecologically) produced, or fed on vegetable diets, or allowed to walk freely, etc. There is a need to develop analytical methods to confirm these claims. There are several good candidate techniques: DNA-based analyses for species identification, protein/proteomics and NMR-based analysis to identify species and processing conditions, and trace element analyses to determine the geographical origin of the material (see review by Martinez et al., 2003b^[1]). All of them need the construction of databases with authentic material so that unknown samples can be analysed under the same conditions as the reference material. The results are then statistically analysed to see whether the unknown matches the authentic sample, or samples suitable as substitutes, or whether it does not match any of the material included in the database (if a new species is used). In order to reduce the likelihood of not identifying the samples it is useful to have as large a database of known reference material as possible.

Species identification is not possible by visual inspection when the morphological characteristics necessary for the identification, such as heads, fins, skin or bones, have been removed during processing. Many methods can be used to uncover fraud (see Table 8) and selection of which will be the most appropriate in a particular case will depend on the nature of the product: whole pieces or comminuted, raw or heat-treated, etc. There may also be a need to determine if a product contains what is claimed on the label or to find out what it really contains. Also, researchers are trying to standardize methods to estimate the amount of each fish species that has been used in fish products (Rehbein and Horstkotte, 2003; Sotelo et al., 2003).

TABLE 8

Information provided by different authentication methods and the molecules targeted

3.3.1 Traditional methods

Traditional species identification is usually based on the analysis of either the proteins or the DNA contained in the product (reviewed by Sotelo et al., 1993; Martinez, 2003; Martinez et al., 2003b). Both protein and DNA-based methods can be subdivided in two groups (Table 8):

i) those that target many loci and produce a fingerprint-like pattern: SDS-PAGE (Piñeiro, et al., 1999a; Mackie et al., 2000]; IEF, (Mackie et al., 2000; Rehbein et al., 1999a; Renon et al., 2001); RAPD (Lee and Chang, 1994; Martinez and Daniélsdóttir, 2000; Martinez and Malmheden Yman, 1999; Partis and Wells, 1996); and

ii) those that target one or a few loci: ELISA-like, immunological techniques, SSCP (Mackie et al., 1999; Rehbein et al., 1999b) and RFLP (Mackie et al., 1999; Meyer et al., 1995; Wolf, Rentsch and Hübner, 1999; Wolf et al., 2000) of PCR amplified sequences.

DNA-based techniques have the advantage that one does not need a standard for each tissue, because almost all the cells in an individual have the same genomic DNA. This is not the case for proteins, since each cell expresses the proteins necessary for its function, which depends on its activity and varies in response to stimuli. The protein pattern in fish muscle is, at least, species-, tissue-, and developmental stage-specific (Martinez et al., 1991).

Which technique is best for product authentication depends mainly on how the sample has been treated. SDS-PAGE or RAPD are easy and relatively cheap to perform, and do not require any previous knowledge of the sample under study (Martinez and Malmheden Yman, 1999; Partis and Wells, 1996; Welsh and McClelland, 1990; Williams et al., 1990). However, they do demand a certain integrity of the proteins or of the genomic DNA. If the product has been subjected to severe processing conditions (such as sterilization in canning) one has to rely on techniques that target small DNA fragments. This is because proteins (Martinez et al., 2000; Sotelo et al., 1993) and DNA (Mackie et al., 1999; Martinez, 1997) are altered so much that they do not render recognizable patterns.

Both nuclear (nDNA) and mitochondrial DNA (mtDNA) regions have been used for species identification. Animal mitochondrial DNA presents inter- as well as intra-specific variability (Moritz, Dowling and Brown, 1987), it is haploid, nonrecombinant DNA, usually maternally inherited (Saavedra, Reyero and Zouros, 1997) and with a mutation rate 5 to 10 times higher than nDNA. Analysis of mtDNA has been proven useful to identify species (Barlett and Davidson, 1991; Bouchon, Souty-Grosset and Raimond, 1994) even in products where the DNA has been severely degraded (Quinteiro et al., 1998). Other alternatives to sequencing are RFLP (Carrera et al., 1998; Meyer et al., 1995) and SSCP analysis of amplified sequences, (shown in Figure 6). This figure also illustrates the need to design different consensus primers depending on the species one wishes to discriminate. While each fish species gave a specific pattern, there were intra-species polymorphisms among the harp seals, and the difference between minke and fin whales was hardly noticeable. These techniques, however, have been optimized for groups of species, as described below, in which case they perform satisfactorily.

DNA target-recognition techniques targeting one or a few loci, are based on the use of PCR for which one has to have some knowledge of the sequence to be amplified in order to design the primers necessary for amplification. Of course, the ideal would be to have species specific sequences for each possible species. When the sequences are unknown, one can use consensus or degenerate primers and once the target is amplified, its sequence has to be confirmed. Although the number of sequences available increases by the day, there are more than 20 000 species of fish used as food, and there are not as yet primers specific for each one. Bartlett and Davidson (1992) list some of the characteristics that a DNA region should fulfil to be a good marker for species identification. Once the target is amplified, confirmation of the sequence can be done by RFLP or SSCP analysis (Mackie et al., 1999; Martinez and Daniélsdóttir, 2000; Meyer et al., 1995; Quinteiro et al., 1998; Wolf, Rentsch and Hübner, 1999; Wolf et al., 2000) or by sequencing (Barlett and Davidson, 1992). The results of species identification based on the recognition of targeted proteins (by using specific antibodies) or DNA sequences are usually positive-negative: either there is a match in which case identity is confirmed, or there is not. Some limitations of these techniques are that they require the development of specific antibodies or primers and that they may render false positive results due to cross-reactivity, cross-hybridization or false positive amplification, with species closely related to that targeted.

Species identification by PCR-RFLP usually requires the use of more than one restriction endonuclease. Which restriction endonuclease will give discriminant fragments depends on the species under analysis (Quinteiro et al., 1998; Wolf, Rentsch and Hübner, 1999; Wolf et al., 2000). The sequence of a 464 bp amplicon of the mitochondrial cytochrome b gene has been examined to choose restriction fragment endonucleases that permit the subsequent successful identification of species of gadoids (Calo-Mata et al., 2003), flatfishes (Sotelo et al., 2001), salmonids (Russell et al., 2000) and raw and processed eel (Rehbein et al., 2002) by the cheaper and easier PCR-RFLP analysis. The same approach, but using a fragment of the mitochondrial DNA control region sequence, has been used to identify species of the Merluccius genus (Quinteiro et al., 2001). Several techniques have been tested for their suitability to identify different species of commercially important cephalopods by Chapela (2003) and Chapela et al. (2002): sequencing of two diagnostic fragments from the mitochondrial cyt b and of the 16S RNA gene were the most reliable and permitted identification of all samples, while RFLP analyses of the discriminant sequenced amplicons was the second best.

Authentication using PCR-SSCP is based on the differences in electrophoretic mobility of each of the strands of a PCR amplicon, differences in mobility reflecting differences in sequence. Again, the suitability of the analysis depends on the suitability of the DNA fragment selected for the amplification. As indicated above for PCR-RFLP, primers should be developed for each group of species, and there may be intra-specific differences that make identification problematic (Martinez and Daniélsdóttir, 2000 and Figure 6). Sequencing is by far the most reliable technique but it is also expensive and time consuming, therefore its routine use by most laboratories is prohibitive at present.

All these target-recognition techniques demand some previous knowledge of the sample under analysis and make the identification procedure longer, more complex and expensive. However, they are the only methods that can be used when the DNA is severely degraded during processing. The reliability of some of these analyses has been confirmed by validation studies among several European laboratories (Hold et al., 2001 a, b).

Sequences of nuclear DNA from the actin gene family have been used for species identification in vertebrates. Actin genes have encoding regions for different isoforms of the protein (isoforms are different molecular forms of functionally related proteins that differ only slightly in their structure), and introns (long and often multiple noncoding intervening sequences of nucleotides contained in the nuclear RNA) that show considerable variation in number and sizes. This has permitted the design of primers complementary to the conserved regions that amplify the variable introns, generating a series of amplicons that, after electrophoresis, produce a fingerprint-like species-specific pattern (Fairbrother et al., 1998; Hopwood et al., 1999). Other nuclear genes used for speciation include regions of the growth hormone gene (Venkatesh and Brenner, 1997) and 5S rRNA gene (Pendas et al., 1994).

Some advantages of fingerprint techniques are that they always produce a result, and false positive identifications are unusual: a “bad” fingerprint (due for example to degradation of the target) is unlike a “good” fingerprint of another species; and once a good fingerprint or a clear sequence are obtained, even if the sample under examination is not contained in the database of references, it is possible to treat the data statistically to evaluate the degree of relatedness between the unknown sample and the reference samples of the collection (Martinez and Malmheden Yman, 1999). On the other hand, identifications based on fingerprints are not useful in case of mixtures (Martinez and Malmheden Yman, 1999), or as mentioned above, if the proteins or DNA are heavily degraded. Fingerprint-based identifications are performed by comparison to reference samples, preferably analysed in the same run as the unknown sample (Martinez, Friis and Seppola, 2001).

Identification of the species from which a pure compound (other than DNA or proteins) is obtained, such as oils, sugars or essences, is usually not possible with the above techniques because either the product contains proteins or DNA as contaminants only or, more likely, they are undetectable. For these products alternative methods must be used: for example in capsules of oils, the analyses of the lipid profile by nuclear magnetic resonance techniques has been shown to be suitable to identify the species used (Aursand et al., 1999).

3.3.2 Quantitative PCR for species identification

Many fishery products consist of mixtures of ingredients and may contain several species mixed in different ratios. Because the total and relative amount of fish and the species may be determinant for the use of the product and the price, many of these products indicate those two data on the labels. Due to the lack of validated methods for the speciation and quantification of fish in foods, the European Union has financed a project to cover the gap (Development of molecular genetic methods for the identification and quantification of fish and seafood. Project QLK1-2000-00476).

The task is not easy, because of the difficulty of first finding suitable species-specific DNA-regions and the problems inherent to quantification, since correct quantification requires that (Rehbein and Horstkotte, 2003; Sotelo et al., 2003):

a) the ratio between target DNA and amount of tissue (g DNA per gram flesh) must be the same for different parts of the fillet and not be influenced by biological factors (sex, age, season, feed availability, stress, etc.);

b) the ratio mentioned in (a) should be known for each of the species to be analysed in mixtures;

c) the procedure for DNA extraction must be very reproducible;

d) the yield of DNA extraction must be the same for all species (ingredients) under examination in mixtures;

e) the effect the processing conditions on the ratio DNA/tissue and on the yield must be known for each of the species and processes (Bauer et al., 2003).

Sotelo et al. (2003) tested the reproducibility of a detection and quantification method based on the Taqman assay for gadoids (Figure 7), and found that some of the main problems lay in the reproducibility of the yield of the DNA extraction: almost a 29 percent variation was found in the yield of DNA in extractions from three individual cod and three replicate samples from each cod. They also found that quantification of cod DNA using the Taqman assay was most reproducible when the amount of DNA used per amplification was between 10-25 ng, whereas lower DNA amounts gave more variable results.

The sequences of the nuclear genes (better suited than mitochondrial genes regarding variability due to biological factors (Battersby and Moyes, 1998) that amplified in a number of species of interest were chosen by Rehbein and Horstkotte (2003) to construct species-specific and consensus primers. These were then tested to determine the gene copy numbers of parvalbumin in Atlantic salmon (Salmo salar), calmoduline and myostatine in saithe (Pollachius virens), and calmodulin in yellow fin tuna (Thunnus albacares). Applications of this technique to baby foods labelled as containing saithe in different amounts revealed that at least 3 of 5 products were mislabelled and in the two others the proportion of saithe indicated on the label was higher than that detected by the analysis.

3.3.3 Determination of geographical origin

To fulfil European legislation, and also to protect endangered fisheries, not only the species but also the location of capture needs to be documented. It is important to note that there is no clear definition of what a “stock” is. Taylor and Dizon (1999) argued that “management units” should be defined on a case-by-case basis and that the objectives of the policy established for the unit should be taken into account early in the design and execution of the scientific approach to preserving the unit. For example, in the case of identifying the origin of samples of marine mammals, Taylor (see page 14 of Dizon et al., 2000) suggested that the most efficient manner to answer the question of “where a sample came from” would be to take a hierarchical approach. First one would identify the species, then the ocean basin (for migratory animals), allocate the sample to a highly distinct population and then, finally, to the stock. This approach demands an exhaustive genetic analysis of each of the segments in the hierarchical level. It requires collection and analysis of a number of samples dependent on how different the different groups being compared are (see page 17 of Dizon et al., 2000).

The techniques suitable for determination of geographic origin have been examined in recent review by Martinez et al., (2003b) and much of the text that follows in this section has that publication as source material. Analysis of the genetic material may also indicate the geographical origin of capture of wild specimens in cases where it is possible to identify genetic markers specific to each geographical area. Martinez and Pastene (1999) described three discriminant RAPD markers to identify minke whales (Balaenoptera acutorostrata) from the North-West Pacific and the North eastern and central Atlantic (Figure 8). Unfortunately, more often than not, it is not possible to find clear-cut markers to allocate the specimens to their location of capture. Thus these authors could not determine in which area of the Atlantic or the Pacific the minke whales had been captured (Martinez et al., 1999) or the origin within the North Atlantic of the Northern shrimp (Pandalus borealis) (Drengstig et al., 2000; Martinez et al., 1997) or Polar cod (Boreogadus saida) using DNA or protein-based analysis (Fevolden, Martinez and Christiansen, 1999). Also, in case of migratory animals, one needs to add a temporal component to the genetic analysis - that needs to be mapped in advance - in order to allocate the individuals to their correct location at each time of the year.

Most genetic studies using either DNA, protein or lipid (Joensen and Grahl-Nielsen, 2003) markers rely on the differential frequency with which the markers are expressed in each population. Allocation of single individuals to their correct stocks can be done by multivariate data analysis techniques and although classification is seldom 100 percent correct, it is usually good enough to permit a very reasonable estimation of the origin, and usually better than classical methods (Aursand et al., 1999; Joensen and Grahl-Nielsen, 2003).

Identification of the origin of farmed specimens may be more difficult than identification of the origin of wild specimens because genetic analysis will not necessarily provide reliable information. Not so long ago, a few countries with a “good” genetic stock could export juvenile fish to many others (this is strictly regulated or forbidden nowadays to avoid spreading of diseases and invasion of the environment by foreign species); and feed can be purchased from international companies, so the composition of the feed does not necessarily reflect what would be natural in the given location. Atlantic salmon (Salmo salar) for example, is currently being cultivated in Norway, Finland, Scotland, Ireland, Faroe Islands, Iceland, Poland, east and west coasts of Canada and United States, Chile and Australia.

Techniques considered more suitable for determination of origin are stable isotope analyses and trace element signature (Table 8). The ratio of the stable isotopes ¹³C/¹²C gives a straightforward answer about the primary photosynthetic metabolism of plant products (O’Leary, 1981), ¹⁵N/¹⁴/N on the type of diet (herbivorous, omnivorous) (Delgado and Garcia, 2001) and on the organic (or not) production of vegetables (Rhodes, 2001), and ¹⁶O/¹⁸O and ²H/¹H are good indicators of environmental conditions (Ziegler et al., 1976). The two main techniques used to determine the isotope ratios of natural products are isotope ratio mass spectrometry (IRMS) and site-specific natural isotope fractionation studied by nuclear magnetic resonance (SNIF-NMR). NMR has the advantage over IRMS that the natural abundance of ²H isotopomers may be precisely and accurately quantified by SNIF-NMR (Martin and Martin, 1991), whereas IRMS only gives a mean value of the deuterium content of a given chemical species. SNIF-NMR is the official method adopted by EU for the authentication of wines (Off. J. Eur Community, 1990). This method is also included in the French standard for canned mackerel (AFNOR NF V45-064 Juin 2002, Conserves appertisées de maquereau - Annexe E) in view of authentication of wine vinegar in marinated products.

Organisms accumulate in their tissues the elements found in the environment in which they live, from the water, food and air. Differences in the isotope distributions of these trace elements in various geographical locations give a different “signature” of isotopes in the organism. Using this technique, it has been possible to distinguish between Atlantic and Mediterranean tuna (Secor et al., 2002), different spawning aggregations of cod (Campana et al., 2002), salmon parr from 14 rivers feeding into the Trondheimsfjord in Norway (Lierhagen, unpublished results, Norwegian Institute for Nature Research, Trondheim) and between oysters from different rearing areas in France (Cardinal et al., 2000). Otholites and/or scales are very good material for this analysis (Campana, 1999; Campana and Thorrold, 2001). One of the most advanced techniques to perform this analysis is called induction coupled plasma mass spectrometry (ICP-MS). The extent to which different processing conditions and packaging materials can affect the trace element signature has not yet been tested.

Lipids may also be good candidates to identify the geographical origin of foods and fish: significant differences in the non-statistical distribution of ²H analysed by NMR, seem to be suitable to classify Atlantic salmon from different sources (Aparicio et al., 1998, Aursand and Axelson, 2001, Aursand, Jørgensen and Grasdalen, 1995a; Aursand, Mabon and Martin, 2000). Also the “fingerprint” that results from the chemical shift position and peak height of ¹³C NMR spectra of lipids has been used to identify the species and origin of purified marine oils (Aursand et al., 1995 a, b). Figure 9 shows the classification of 112 different oil samples into 25 different classes using fuzzy classification (Bezdek and Castelaz, 1977) of ¹³C NMR spectra (data from the unpublished results of Dr Aursand and Dr Axelson). The lipid profiles obtained by gas chromatography, IRMS, and high-resolution ²H SNIF-NMR spectroscopy have been used to classify different kinds of fish oils and lipids extracted from muscle of wild (Norway) and farmed (Norway and Scotland) salmon (Aursand, Mabon and Martin, 2000). A statistical analysis of the fatty acid compositions, overall ²H and ¹³C isotope ratios, and molar fractions of the isotopomeric deuterium clusters was carried out to select the most efficient variables for distinguishing the different groups of salmons and fishes studied. A canonical discriminant analysis with eight mixed variables (four fatty acid compositions, three deuterium molar fractions, and the overall (D/H)tot isotope ratio of fish oils) gave 100 percent correct classification of the oils. NMR techniques have also been applied to estimate stress and freshness of the fish (Erikson, Beyer and Sigholt, 1997; Sitter et al., 1999) and whether the fish (cod) had been frozen (Howell et al., 1996).

3.3.4 Proteome analysis

By analogy to the term “genome” (the complete set of the genetic material of a cell and/or organism), the term “proteome” refers to the entire protein complement encoded by an organism’s or a cell’s DNA (Nelson and Cox, 2000). The proteome is dictated by the genome and is therefore species-specific. As opposed to the genome (which is mostly identical regardless of the tissue, age, physiological status etc., of the organism), the proteome is highly dynamic: the proteins expressed by a genome and the post translational modifications suffered by them vary depending on stimuli received, age, sex, diet, physiological status, etc. Because of this variability, the proteome is defined as “the proteins present in one sample (tissue, organism, cell culture) at a certain point in time” (Kvasnicka, 2003).

Proteomics is the set of tools used to characterize the proteome: roughly, all the proteins have to be separated first and then each one of them is digested and the fragments analysed usually by mass spectrometry (MS); then either the peptide fingerprint of the protein or the amino acid sequence of some of its fragments are compared to known DNA or protein sequences from databases to identify the protein. The amount of data generated by proteomics and genomics is such that it requires the use of significant computer power and development of software for identifications and searches. This new field is called “bioinformatics”.

The characterization of thousands of proteins simultaneously was made possible thanks to the development of the two-dimensional electrophoresis (2-DE) by O’Farrell (1975). This technique involves separating the proteins contained in an extract, first according to their isolectric point, and secondly according to their molecular mass in a solid matrix of polyacrylamide. The proteins are then stained (see Patton, 2002 for a review on detection methods) and one can see differences between different samples by differences in the protein map (Figure 10 shows a two-dimensional gel). Identification of the proteins is then made possible by cutting the protein spots from the gel, performing amino acid sequencing and comparing it with sequences of known proteins or the estimated amino acid sequences of known gene DNA sequences. This procedure was initially technically very demanding, requiring highly trained personnel, and only a few samples could be analysed a day. Later improvements have made it easier but still many researchers are looking for alternative procedures with fewer limitations. Liquid chromatography (LC) and high performance (pressure) liquid chromatogray (HPLC) seem to be very promising. A review by Kvasnicka (2003) examining the new proteomic techniques and their application to nutritionally relevant proteins has recently been published.

Piñeiro et al. (2003) have reviewed the numerous applications of proteomics to the investigation of seafood and marine products, which include: the detection of shellfish toxins, allergens, antifreeze glycoproteins (Draisci, et al., 1995; Hess et al., 2001), identification of the species (Martinez and Friis, 2003; Piñeiro et al., 1998, 1999b, 2000, 2001, 2003), developmental stage (Martinez et al., 1991; Martinez and Christiansen, 1994), characterize post-mortem changes in fish (Kjærsgård and Jessen, 2003; Verrez-Bagnis et al., 2001) and crustacean muscles (Martinez et al., 2001b), examination of some processing conditions such as freezing/thawing (Jessen, 1996), fermentations (Morzel et al., 2000), surimi manufacture (Martinez et al., 1992), and presence of contaminants in the environment where seafood is cultivated (Rodríguez-Ortega et al., 2003; Shepard and Bradley; 2000; Shepard et al., 2000).

3.3.5 Microarrays

The first microarrays, also known as “chips”, contained DNA; nowadays there are protein and DNA microarrays.

DNA microarrays are miniature high density arrays that can have from a few hundred to several hundred thousand oligonucleotides (probes) capable of taking part in hybridization reactions (Figure 11). The probes are bound to a solid support (nylon membrane or glass). The sample to be analysed is labelled (usually with a fluorophore) and applied to the array under hybridization conditions (see Southern 1997, 2000; Southern and Elder, 1995; Southern and Maskos, 1995). Although the terms DNA micro-arrays and DNA chips are sometimes used complementarily, Jordan (1998) made a distinction. He used the term micro-array to refer to sets of DNA targets deposited on a solid support (generally a glass microscope slide) with a spot density of several hundred individual spots per cm², and the term DNA chip (or oligonucleotide chip) to refer to large sets of oligonucleotides synthesized in situ, usually at much higher densities.

The arrays are usually created by computer controlled robots that spot the probes onto precise locations on the support. The sample under analysis is labelled with a fluorogenic dye and added to the array, where it hybridises to its complementary probe if it is present in the array. After UV irradiation, the bound sequences fluoresce and their intensity is measured by a detector. In addition to three patents of Southern (Southern 1997, 2000; Southern and Maskos, 1995), Gerhold, Rushmore and Caskey (1999), Jordan (1998) and Ramsay (1998) give reviews and a description of the technology, operating principles and applications. These include: detection of single nucleotide polymorphisms, linkage analysis, gene expression and the effects of stress, drugs or diseases on gene expression. DNA arrays have also been used for the simultaneous species identification, genotyping and assessment of antibiotic (rifampin) resistance of Mycobacterium spp. (Gingeras et al., 1998; Troesch et al., 1999).

In these arrays or chips, the hybridization is carried out under conditions where the reaction rates and stringency are controlled by the temperature, salt concentration of the hybridization solutions and washes, and the concentration of the target DNA. A later development in this technology exploits the use of electric fields (active microelectronic chips) to transport and concentrate the negatively charged nucleic acid molecules to the selected (positively charged) microlocations on the array (Heller, Forster and Tu, 2000).

Protein arrays contain small spots of proteins that are immobilized onto silicon-based substrates, typically glass. As with DNA arrays, protein arrays can be used to screen crude mixtures and examine protein-protein or protein-ligand interactions (see Kvasnicka, 2003). As opposed to the complicated proteomics techniques, these arrays can be relatively cheap, easy to use and fast. However, development of suitable protein arrays do require the prior use of proteomic techniques, either in standard or in “micro” formats (Kvasnicka, 2003), to identify the protein or peptide markers that are suitable for each application. Whether for protein or DNA analyses, use of microarrays requires very small amounts (µl) of sample

3.3.6 Automated procedures

The most valuable item in an analysis is usually the sample to be analysed. Yet, mixing up samples or losing them while carrying out the procedures, as well as contamination and cross-contamination, are not uncommon events. The development of equipment and protocols for automated sample preparation - preparation of broths for microbial analyses, as well as DNA or protein extractions, and setting up PCR reactions - are important, especially if a certain number of samples must be processed every day. Some companies have realized the importance of these steps and such equipment exists, but due to the type of market the equipment has been designed to suit clinical laboratories processing many liquid samples (blood, urine) each day. The development of similar equipment for food control laboratories lags behind.

In the case of microbial analysis, it would be very convenient to have some method to separate the micro-organisms from the rest of the broth or food matrix. Advances in the development of microelectrode structures have led to new techniques for the dielectrophoretic characterization and sorting of cells, micro-organisms and other bioparticles using non-uniform alternating current (AC) electric fields. Dielectrophoresis is the lateral motion induced to particles by non-uniform electric fields. Applications of dielectrophoresis have included the selective spatial manipulation and separation of mixtures of bacteria, including Bacillus subtilis, E. coli and Micrococcus luteus, viable and unviable cells, cancerous and normal cells, and red and white blood cells as well as viruses and macromolecules such as DNA and proteins (Markx et al., 1996; Pethig and Markx, 1997). Although the technique has the potential of being incorporated into “laboratory-on-a-chip” devices, this property has not yet been exploited commercially to develop equipment suitable to handle food samples.

Many of the newest methods for protein analyses require the use of electrophoresis to separate the products under analysis and to visualize the results. Electrophoresis is relatively slow, it requires additional steps for visualising the results (staining and destaining, developing of films) and it is highly dependent on the matrix used. Variables such as the quality of the reagents used to prepare the matrix and the buffers, the equipment used, the temperature of the run and the ability of the person performing it have a critical effect on the result obtained. An example of a “bad” run - a distorted gel - is shown in Figure 3. Newer techniques to eliminate this step and make it more reproducible and faster are also needed, and indeed Binz et al. (1999) presented a highly automated method to generate fully annotated 2-DE maps. Using a parallel process, all proteins of a 2-DE are first simultaneously digested proteolytically and electro-transferred onto a poly (vinylidene difluoride) membrane. The membrane is then directly scanned by matrix-assisted laser desorption/ionization - time of flight - mass spectrometry (MALDI-TOF MS); after automated protein identification from the peptide mass fingerprints obtained using PeptIdent software (http://www.expasy.ch/tools/peptident.html), the method generates a fully annotated 2-D map on-line.

Other researchers prefer to use methods where gel electrophoresis is substituted by liquid in-line techniques (see Kvasnicka, 2003). For example, the use of multidimensional HPLC for protein analysis (tandem or multiple LC combined with tandem MS) allows for easy automation and permits the separation of low-abundance and membrane proteins. Developments of techniques based on the use of microfluidic systems permit the development of lab-on-a-chip approaches where the liquid samples to be analyzed move thorough the inside of tiny microchannels using combinations of hydrophobic surface modifications, electrokinetic transport and air pressure. The widespread interest in micro total analysis systems has resulted in studies to avoid technical problems, such as liquid slip, and optimise the procedures (Yang and Kwok, 2003), as well as efforts to develop devices in cheaper polymer materials as an alternative to expensive glass and silicon devices such as polydimethylsiloxane (PDMS) (Wang et al., 2003).

It is foreseeable that in the future, the whole process of a sample through analysis will be automated. Thus, from one sample, sub-samples can be taken for authentication and for microbiological analyses. Tailor-made DNA or protein micro-arrays combined with lab-on-a-chip techniques can be used to detect multiple targets (species, species and strains of bacteria, genes for resistance to antibiotics and disinfectants, etc.) in one step. Optimally, all these steps should be carried out in a self-contained apparatus. The results from these analyses have to be interpreted (intensities have to be converted to amounts for each of the targets) by the computer software provided with the equipment. With new demands for a full traceability chain for each batch, the information provided by these analyses should be linked to the history of the batch in order to detect any problems or to certify that the batch fulfilled all the required standards at the time of the analysis.