CHAPTER 1: The Role of DNA Analysis in the Determination of Food Authenticity

Friedrich Accum was a German chemist who came to London in 1793 and established himself as an analyst. His interests included keeping processed food free from dangerous additives. In 1820 he published a best-selling book entitled ‘A Treatise on Adulterations of Food and Culinary Poisons’.¹  In this book he describes practices such as colouring red cheese and confectionary with red lead and mercuric sulphide, using strychnine instead of hops in the production of beer, whitening bread flour with alum and chalk and extending bread loaves with plaster of Paris and sawdust. Other practices included boiling spent tea leaves with Prussian Blue (ferric ferrocyanide) and sheep dung, which were dried and re-sold.

Arthur Hill Hassall (a London physician) began a major investigation into food adulteration in the 1850s. One of his early successes used microscopy to demonstrate the adulteration of coffee with chicory. Hassall's work, together with Henry Letheby, the Medical Officer of Health for London, led to the introduction of the Adulteration of Food and Drugs Act of 1860. This Act was revised in 1872 and again in 1875. The Sale of Food and Drugs Act 1875 is widely regarded as a turning point in the regulation of food, introducing key concepts such as that food must be of the ‘nature’ or ‘substance’ or ‘quality’ demanded by the purchaser. It included, as a duty of local government, the appointment of a certain type of scientist, the Public Analyst – a key Hassall recommendation – to provide the underpinning analytical data and its interpretation for the enforcement of the provisions of the Act. The Society of Public Analysts was formed (now known as the Association of Public Analysts). Today, one of the key tasks of public analysts in the UK and Ireland is to ensure the safety and correct description of food by testing for compliance with legislation.

As described in Chapter 10, legislation that relates to the authenticity of food remains in place today. In the UK, the Food Safety Act 1990 prohibits ‘falsely or misleadingly describing or presenting food’. Regulation (EU) No.1169/2011 of the European Parliament and of the Council of 25 October 2011 on the provision of food information to consumers, (implemented in the UK as the Food Information Regulations 2014, No.1855) as amended and with devolved equivalents specifies what information must be given with marketed food. For pre-packed food the required information includes a list of the individual components, the amounts present (quantitative declaration of ingredients or QUID) and, where appropriate, the country of origin of key ingredients. Inherent within this regulation is that, with very limited exceptions, pre-packed food should not include any undeclared ingredients. Despite this legislation, food adulteration and other forms of food fraud continues to exist. INTERPOL and Europol have undertaken joint operations targeting counterfeit and sub-standard foodstuff and beverages. For example, operation OPSON V in 2015, involving 57 countries, where the operation seized over 10 000 tonnes of food and one million litres of drink which was either counterfeit or sub-standard.

Consumers expect that the food they buy is labelled correctly but, as the example above shows, food fraud is prevalent despite the existence of relevant legislation. Food fraud can broadly take three forms: adulteration, substitution and mis-description.

Adulteration includes the addition of undeclared ingredients. Examples include adding horsemeat or offal to beef in processed foods, adding illegal colourants to improve the appearance of food and adding water to frozen food to increase its weight. Other examples of food adulteration are given in Box 1.1.

Substitution involves replacement of one ingredient by a similar or cheaper one. Examples include using whiting or pollack in place of cod, bonito in place of tuna or sea trout to replace salmon. A survey by Oceana found this type of fraud to be widespread in restaurants in the USA and Europe.²  Similar studies aimed specifically at sushi restaurants found substitution varying from 10% in the UK to 47% in the USA.^3,4 

Misdescription of food takes several forms and usually relates to the method of production, the geographical origin of the food or the amounts of key ingredients used. Some representative examples are given in Table 1.1.

In the USA, food fraud is treated primarily as a food safety issue. In contrast, in Europe food fraud is viewed as a breach of food labelling law with food safety being covered by separate legislation. Nevertheless, there is a clear link between food fraud and food safety. Examples include the US Oceana surveys,²  where most of the fish labelled as tuna was in fact escolar, which can cause digestive problems. Another example of a safety issue related to adulteration is the use of known carcinogens (illegal dyes), such as Sudan 1, to improve food colouration and increase the price of spices such as chilli powder.

There are a number of drivers of food fraud. Two key ones are price pressures on food suppliers and criminal activity. Low profit margins for food manufacturers create cost pressures if raw materials or utilities increase in price. This can potentially create opportunities for food fraud by reducing costs via adulteration or substitution of key ingredients. Criminal involvement usually is associated with high value or premium products where huge profits can be made by passing off a cheaply produced or inferior quality product as a superior version. Such criminal activity operates at all levels of food production; and is attractive due to its perceived low penalties, for example, compared with drugs trafficking.¹ Further discussion on food fraud is provided in Chapter 14.

There are five main challenges in developing methods to determine the authenticity of food. Each generates a requirement that the methods are fit for purpose. The first challenge is the choice of analytical method. The range of misdescription and fraud described in the previous section means that no single methodology will suffice. As discussed elsewhere in this book, it may be necessary to use multiple methods.

The second challenge is finding a marker or markers that characterises the food, one of its ingredients, the adulterants in question, or its processing, production or geographical origin. The marker has to be specific, its natural variation must be limited and well-characterised and be measured accurately.

The third challenge relates to the variety of matrices that exist in foods. No two processed foods contain exactly the same ingredients. A method developed for one food may not be suitable for use with another. If processed foods are cooked, any markers of interest may be degraded or destroyed completely.

The fourth challenge is that most investigations of food adulteration are linked to a legal requirement, standard or guidance. As such, the interpretation of the results must be made in the light of analytical uncertainty, natural variation and any tolerance permitted by the requirements. That is, the conclusion reached must be beyond reasonable doubt.

The final challenge relates to the requirement for authentic samples and/or certified reference materials for the development and evaluation of the method. The difficulties in obtaining authentic samples cannot be over-emphasised. Often, once sourced, there is no central repository for their maintenance and supply to analysts. Further discussion on drivers and challenges for food fraud adulteration and its determination is provided in Chapter 6.

All cellular organisms contain DNA. The differences between organisms i.e. anatomical, physiological etc. ultimately resides in their DNA sequence. The more evolutionarily distant two species are, the greater will be the differences in their DNA sequences. However, even when two individuals are closely related, e.g. family members, there are sufficient differences at the DNA sequence level for each individual to be recognised. This is the basis of DNA profiling used in police and paternity investigations and which came to prominence in the 1980s.

In the early 1990s the UK Ministry of Agriculture, Fisheries and Food (MAFF) began funding research to develop analytical methods to detect food fraud, which included the application of forensic DNA profiling to determine the authenticity of food. DNA analysis is particularly suited to qualitative analysis to answer yes/no questions e.g. has there been substitution of one species with another, which formed the focus of much of the early DNA-based methods work. As explored in Chapters 6 and 7, public analysts were not overly familiar with the diversity of emerging new DNA techniques for food testing. This issue was resolved when the Agilent lab-on-a-chip analyser became available resulting in the conversion of these DNA techniques to this easier to use format (Chapter 6).

More recently, qualitative analysis has been facilitated by the development of DNA barcoding (www.barcodeoflife.org). Barcoding provided a way of distinguishing and identifying species with a short, standardized gene sequence.⁵  A 648 bp region of the mitochondrial cytochrome-C oxidase-1 (CO1) gene was proposed. This proposal has been adopted internationally for the identification of many animal species. However, in plants there is insufficient variation in this gene to be useful. For plant species identification, two gene regions in chloroplasts, matK and rbcL, are often used, although there is still no agreement on a universal barcode to use for plants.

Police and paternity forensic work and food authenticity determination differ because for the latter there is often a requirement for quantitative analysis to authenticate foods. This requirement stems from the need to distinguish between adventitious contamination and deliberate fraud. Historically it was accepted that, as a rule of thumb, undeclared species at less than the 5% (w/w) level was adventitious contamination.

Following the 2013 horsemeat incident,⁶  the UK Food Standards Agency set an upper limit of 1% (w/w) for undeclared ingredients with a requirement for explanation if an undeclared ingredient was found to be present at greater than 0.1% (w/w). However, there are exceptions. For example, for EU approved Genetically Modified Organisms (GMOs), food and feed products that contain <0.9% (w/w) undeclared genetically modified material on an ingredient by ingredient basis do not need to be labelled as this is considered adventitious or technically unavoidable contamination. Also, in Europe, Basmati rice can be imported tariff-free provided it contains less than 7% unapproved varieties.

As described in Chapters 6, 9 and 20, quantitative analysis of DNA is based upon the polymerase chain reaction (PCR) which is a method of DNA amplification. Quantitative PCR (qPCR), also known as real-time PCR, measures the copy number of the marker gene (for the ingredient or adulterant) and compares it to the copy number of a ‘normalising’ gene. This ratio is compared to a standard mixture to give the relative percentage of the ingredient or adulterant. Although the concept of qPCR is very simple, reducing it to practice with the necessary precision and accuracy is often challenging, particularly with many food matrices.

Digital PCR (dPCR) is a technique which facilitates absolute single molecule detection, without the need for a standard curve. Digital PCR offers a number of potential advantages for quantitation, including permitting good precision estimates through sample partitioning, enrichment of minority targets and being less subject to partial inhibition, and is more fully described in Chapter 5.

There are a number of ways in which a particular gene sequence can differ between different species. The commonest one is a change in a single base-pair, known as a single nucleotide polymorphism (SNP). This usually arises as a result of mutation. If a SNP occurs in a DNA sequence that acts as a recognition site for a restriction endonuclease, it is known as a restriction fragment length polymorphism or RFLP. SNPs arise very infrequently in any particular gene and are most useful in distinguishing unrelated species, e.g. beef, lamb, pork and chicken. Changes in the number of simple sequence repeats are known as simple sequence length polymorphisms (SSLPs) or microsatellites, and usually occur as a result of mistakes in DNA replication. SSLPs occur with a much higher frequency than SNPs and, as with human DNA profiling, are particularly useful in distinguishing closely related organisms such as domestic pig versus wild boar⁷  and varieties or cultivars within a species e.g. rice varieties.⁸ 

A generic method of DNA analysis is as follows: DNA is extracted from the test sample and purified. A selected region of the DNA genome (target gene or ‘marker’) is amplified using the PCR and the resulting ‘amplicon’ digested with a restriction endonuclease. The resulting DNA fragments are analysed using separation by agarose gel electrophoresis. Examples applying this method are listed in Table 1.2. The application of this technology was facilitated by the commercial availability of a microfluidic lab-on-a-chip system.^9–11  This permits 12 post-PCR samples to undergo capillary electrophoretic separation, staining and sizing on a disposable chip less than 3 cm square. However, to use this method it is essential that authentic reference material is available.

One disadvantage of the lab-on-a-chip system is that it does not have the resolving power to separate DNA fragments differing by a single repeat unit, i.e. 3–4 base pairs. The alternative is to replace SSLPs as markers with insertions or deletions of DNA of a size to be resolvable. This was the approach taken by Steele et al.¹²  to differentiate Basmati rice varieties using the lab-on-a-chip system (as described in Chapter 19). An alternative approach to the lab-on-a-chip system is the LI-COR DNA analyser. This offers qualitative and quantitative improvements over conventional gel assays and can be used with SSLPs.¹⁶ 

An alternative approach to gel electrophoresis for species identification is to amplify the CO1 gene (for animal products) or the matK and rbcL genes (for plant materials) using the PCR; and determine the entire DNA sequence from the resulting amplicons. The obtained sequence(s) are compared with those in the Barcode of Life Database (BOLD) or other DNA databases. As far as fish species identification is concerned, Lendi et al.¹⁷  have shown that the BOLD constitutes robust and reliable datasets, unlike other DNA sequence repositories. Nevertheless, it is good practice to use reference material of known authenticity. One advantage of this Barcode of Life sequencing approach is that it can detect adulterant species whose presence is not suspected¹⁸  as the approach is a non-targeted technique. For example, Ortola-Vidal et al.¹⁹  used this method to detect as little as 2% (w/w) rhubarb in raspberry yoghurt. As described elsewhere in this book, more significantly, it can detect adulteration of prime meats (beef, lamb, pork) with unrelated species, such as donkey, even in processed meats such as salami. The technique has been particularly useful in identifying mislabelling of fish where potentially there are hundreds of different species that can be substituted.

Earlier, it was noted that SSLPs are particularly useful in differentiating closely related species. However, there are some instances (cattle and pig breeds) where differentiation can only be achieved by using a very large number of markers. Here the markers of choice are SNPs. However, analysis of these SNPs is often conducted using microarrays rather than gel electrophoresis.²⁰ 

Certain cured meats only produced in specific geographic regions are made with meat from particular breeds of animal (usually pigs). A method has been developed to verify the origin of beef or pork from 12 breeds of cattle or 14 breeds of pigs that are considered as specialties in the UK.²⁰  This method is based on analysis of a large number (96) of SNPs and the data generated can be used to answer two questions: ‘Is this meat from breed X?’ and ‘From which breed of cattle or pig has this meat come?’ A positive answer will be obtained if the SNP profile is available for the breed of interest. Thus, it will be possible to determine if the meat present is from Landrace pigs (e.g. Serrano ham) but not Iberico pigs (e.g. Iberico ham). Similarly, ‘Vitellone dell’ Appennino Centrale’ beef should be produced from only three cattle breeds: Chianina, Romagnola and Marchigiana. Potentially, if a more complete SNP database was generated that included all the breeds farmed in the UK and the EU then it might be possible to use the method more generally for country of origin determination. However, this would depend on having information about which breeds are farmed in which countries and the availability of authentic samples or reference materials.

The analysis of pig and cattle breeds described above (and in Chapter 6) involved analysis of panels of 96 SNPs in expensive microarray formats. A potential alternative approach is the use of SINE-based markers. SINES are short interspersed nuclear elements, which are dispersed repetitive sequences present in many higher eukaryotic genomes. Wenke et al.²¹  have developed highly informative primer sets for PCR amplification that can be used to differentiate different potato varieties. The amplified fragments are known as inter-SINE amplified polymorphisms (ISAPs) and can be separated by conventional electrophoresis or analysed using a capillary sequencer.

Animals can exist in distinct populations with little or no interbreeding. As a result, these populations begin to drift genetically, such that DNA sequence differences are detectable. For example, it has been found that Hereford cattle in the UK were genetically different from Hereford cattle in the USA, even though the two populations had common ancestors (Rob Ogden, personal communication). Topographical barriers (e.g. mountains, major rivers) or extensive urbanisation may prevent different groups of animals mixing. Thus, it is likely that deer and some game birds in Great Britain are genetically different from those in Ireland or mainland Europe. These genetic differences could be used to facilitate authentication of country of origin labelling.²²  A similar approach is being used to identify the fishing grounds from which non-migratory fish were sourced.^23,24 

Metagenomics is a technique whereby DNA is extracted from environmental samples and the 16S rRNA genes are sequenced. Bioinformatics techniques are used to identify the microbial species from which the DNA was derived, including species that never have been cultivated in the laboratory. The technique enables a profile of the total microbial flora in a sample.²⁵  Since the microbial flora associated with a sample will reflect the environment from which it was derived, metagenomics is potentially a useful technique for verifying country of origin labelling.

Metabarcoding is a technique which combines next generating sequencing with DNA barcoding. This facilitates next generation DNA sequencing of universal genetic markers together with a reference barcode database for taxonomic identification. This metabarcoding technique has been successfully applied for the simultaneous detection of animal and plant species present as ingredients in a range of food supplements and traditional medicines.²⁶  Next generation sequencing is recognised as a powerful but new and emerging technology for food testing and is further discussed in Chapter 8.

The DNA-based authenticity methods described in this book are likely to be used in two ways. Food manufacturing companies use analytical methods to support traceability and quality control procedures. Chapter 24 provides more detail on commercial DNA testing. Certified analytical laboratories may use these techniques to analyse samples provided by enforcement agencies to verify food law compliance. The second category has an absolute requirement for fitness for purpose because the results obtained may provide evidence in a court of law (Chapter 10). The detailed considerations associated with fitness for purpose of molecular biology methods are discussed further by Burns et al.²⁷  and an overview is presented in Chapter 7. Broader aspects around fitness for purpose, including standardisation of DNA-based methods and collaborative trials, are discussed in Chapters 20 and 21.

The analysis of DNA now is used widely in the determination of food authenticity. The rest of this book is devoted to specific issues associated with its use and the suitability of the different markers used in various authenticity applications. However, as noted in Chapter 6, DNA analysis is only one component of the analytical toolbox needed to verify food authenticity.