Food authenticity testing part 1: The role of analysis

February 2019

This is a fundamental re-write of the IFST’s “Food authenticity testing” Information Statement. It has been split into two parts. This statement now covers the role of analytical testing within the context of an overall supply chain assurance strategy. It describes where testing can and cannot be used, and highlights generic issues relating to interpreting food authenticity testing results.

The description of specific analytical techniques, their applications, strengths and weaknesses have been moved to a separate information statement “Food authenticity testing part 2: Analytical techniques”.

Executive summary

Analytical testing is a valuable tool in the armoury to assure food authenticity but cannot be used to identify every type of food fraud.  It is only one part of an overall strategy to mitigate fraud risk.

Many modern tests are based upon comparing a pattern of measured values in the test sample with patterns from a database of authentic samples. Interpretation is highly dependent on the robustness of the database, and whether it includes all possible authentic variables and sample types. This information may not be released by the laboratory.  Interpretation of results is rarely clear-cut, and analytical results are often used to inform and target further investigation (such as unannounced audits or mass-balance checks) rather than for making a compliance decision.

Food fraud

There are multiple organisations working on an internationally accepted, and legally enforceable, definition of Food Fraud. The 2018 working draft of the Codex Committee on Food Import and Export Inspection and Certification Systems is typical: “Food fraud includes adulteration, deliberate and intentional substitution, dilution, simulation, tampering, counterfeiting, or misrepresentation of food, food ingredients, or food packaging; or false or misleading statements made about a product for economic gain”.  There is debate whether the legal scope should be expanded wider than “…. deliberate and intentional…” and wider than “…. for economic gain.”

The risk of food fraud

Adulteration and mislabelling of food has been known since biblical times.  Fraud goes back as far as when food is traded – ancient Greeks had laws on adulteration of cereals and fats.  Whenever there is a price premium between two ostensibly similar products, or a downward pressure on prices, then there is scope for criminality.  The risk and potential scale have increased in recent years with the complexity of modern supply chains, high value ingredients, and the proliferation of premium-labelled variants of food and drink types.  A recent analysis estimated that fraud accounted for 5 to 25% of all globally reported food safety incidents[1].

In the simplest case, one (safe and legal) food is misrepresented as a more expensive variety. There is an economic and reputational risk to any defrauded trader, but no direct risk to the safety of the consumer. Nobody will be harmed by eating conventional carrots mis-sold as organically produced, or a ready-meal containing a different variety of potato than the manufacturer believed they were buying. But at the other end of a spectrum, the health risk can be severe if a non-food-grade ingredient is added to disguise the quality or variety of a product.  Examples are adding melamine to increase the apparent protein content of milk powder, or Sudan dyes to disguise vegetable oils as palm oil.  Even where there is no direct health risk, it is unlikely that such an unscrupulous trader will diligently be following all other food safety and hygiene rules.  The UK’s stated regulatory approach is therefore to take a zero-tolerance approach to any cases of mis-selling[2], or else food safety incidents will inevitably follow.

The role of analytical testing in a food fraud defence strategy

Food manufacturers and retailers should assess their vulnerability to fraud as part of their routine risk assessments[3] and put risk mitigation steps in place as appropriate. 

The only way for the food industry to guard against being the victim of fraud is robust supply chain defence policies; short and transparent supply chains, financial audits, mass-balance checks and effective whistle-blower procedures. This is often supported by an analytical testing programme, whether designed to directly detect issues, to target further investigation and audit, or purely as a deterrent.  Analytical testing is targeted with the help of data-sharing and early warning systems.  Testing is used as a spot-check to verify that control systems and certification are effective and trustworthy.  It is not in itself a control or certification system.  Typically, food manufacturers will use risk-based prioritisation to conduct unannounced analytical spot-tests on their raw materials, to check that they are exactly what they purport to be.

It is important that manufacturers and retailers then have a procedure in place for acting upon analytical results. Unlike testing for chemical contaminants, the interpretation of results from food authenticity tests will often be ambiguous or have a high degree of uncertainty. There is little point in a manufacturer commissioning a test if they do not know how they will deal with “suggestive, but not conclusive” results.

Scope of analytical testing for detecting different categories types of food fraud

Detecting food fraud or verifying whether the information and voluntary claims accompanying the food are correct, involves analytical tests to examine the composition both qualitatively and quantitatively, processing, geographical origin, and compliance with certification systems. This means that a very wide range of methods have been developed for this purpose, which are often collectively termed “food forensics”.

Many such methods rely upon comparing the test sample with a reference database that contains “normal” or “authentic” samples. These give a probabilistic, rather than a definitive, result. If the test sample differs from the reference database then there may be a range of hypotheses as to why: both fraudulent and innocent reasons. For example, the 14N/15N ratio in a vegetable may differ from the reference database because an organically-certified grower has used a novel, but permitted, seaweed-based fertiliser, or it may differ because they have illegally used a synthetic (mineral) fertiliser. Thus, analytical testing will not always give a definitive result upon which an accept/reject verdict can be made (Figure 1).

Descriptions such as “organic” refer to a legally defined system, and it is not realistic to have one or several methods of analysis which cover this whole system, only inspection and audit can do this. The best that can be achieved analytically is to verify one aspect of the system e.g. soil fertilisation, antibiotic use or plant metabolites as one marker for verifying the use of the term “organic”.

This visualisation echoes the Seven Deadly Sins of food fraud defined by Spink and Moyer[4]. Not all of these sins can be detected by analysis, and of those that can there is not always the luxury of an unambiguous verdict.

Most analytical research and development is focussed upon increasing the certainty of the probabilistic tests; developing analytical techniques, building reference databases of both legitimate and fraudulent products, and developing statistical pattern-recognition software. This area encompasses most of the “up-labelling” fraud risks of current concern such as mislabelling of product grade, species or variety, organically grown produce, country of origin, or artisan production method.

Analytical approaches: fundamental classifications

Irrespective of the analytical technique, there are some fundamental ways of classifying the analytical approach. Many of the techniques can be used for multiple approaches, depending on how they are configured. Typical ways of classifying the approach are:

  • targeted vs untargeted analysis
  • specific analyte(s) vs Multi-Variate Analysis (MVA), or
  • laboratory vs point-of-use testing.
Targeted vs untargeted analysis

Traditionally, analysts have needed to pre-define what they are seeking to measure.  This is “targeted” analysis.  In the field of food fraud, examples are:

  • testing for a specific adulterant (e.g. melamine in milk powder, chicory in soluble coffee powder)
  • testing for a known marker that is characteristic of a particular grade of product (e.g. UV absorbance ratios and fat ratios to characterise Extra Virgin Olive Oil), or
  • testing for a pre-defined section of DNA (e.g. use of specific primers to amplify DNA from a specific species of meat).

Targeted analysis also includes tests where ratios of known markers are used in a more probabilistic manner; when there are no fixed tolerances or thresholds, but there is a degree of knowledge about what is “typical” for certain product types (e.g. measuring the 14N/15N isotope ratio in vegetables labelled as organically-grown to try and infer the mis-use of a synthetic (mineral) fertiliser). 

Targeted analysis tends to be much more sensitive than untargeted, because instruments and techniques can be tuned and optimised for the specific analyte(s) sought.  The obvious limitation with targeted analysis is that if an issue is not sought then it will not be found.  It is always reactive.

“Untargeted” can be used to describe situations where there is still a pre-defined list of parameters but it is so large that, for all intents and purposes, the application is universal.  An example is the use of universal primers for DNA amplification, sequencing and matching against a comprehensive reference database for species identification.

 In its purest form, untargeted analysis has no pre-defined list of test parameters.  All that is known is the pattern of results.  Multiple data points are collected from the sample.  It may not be known which individual parameters or analytes are being measured or what they indicate.  Examples are measuring changes in the complex patterns of proteins, metabolites or genes in a sample, not all of which may be identified, often referred to as “-omics” techniques (e.g. genomics, proteomics, metabolomics).  But there are many other examples, such as measuring complex fat profiles in fish or meat, or the intensity of each (unidentified) peak in the complex Nuclear Magnetic Resonance (NMR) spectrum of an alcoholic drink, or the intensity of each (unidentified) peak in the complex mass spectrum of a dried herb.  In all cases, data assessment involves Multi-Variate Analysis (see Section 2.2) and comparison with extensive reference databases.  For example, an MVA reference database of the fat profile of cod might reveal a pattern of statistical clusters, each associated with fish of the same species but from a different catch area.  If the test sample fell within one of these clusters then it might be inferred that it was also from this catch area.

It is important, in untargeted analysis, to keep a clear distinction between the analytical result and the interpretation of the result. This is particularly critical for laboratory accreditation. Accreditation for “testing” and for “opinions and interpretations” are very different and separate processes. Typically, a laboratory will be accredited for “testing” but only a named expert can be accredited for “opinions and interpretations”. Even a named expert can have his interpretation challenged in a Court of Law so it is not necessarily unequivocal.  Laboratories will take care on their written reports not to stray beyond their accredited scope by commenting on interpretation, and this can mean that customers are unaware of caveats and are left with the false impression that the interpretation of the result is clear-cut.

Untargeted analysis lends itself to spectral techniques where data over an entire signal range is collected: where there is no pre-selection of data. Techniques such as mass spectrometry (in full scan mode), NMR, and spectral imaging using any or all of the infra-red, near infra-red, visible or ultra-violet light ranges.

Specific analyte(s) vs multi-variate analysis (MVA)

MVA is the basis behind all non-targeted approaches[5], but may also be used for targeted analysis if multiple pre-defined parameters are measured; where no individual parameter or ratio is a marker for the result, but the overall pattern gives an indication of the result.  An example of targeted MVA is stable isotope ratio mass spectrometry where the isotopic ratios of four or five different natural elements are plotted to give an indication of geographic origin. 

The pre-requisite for MVA is the construction of a database of results from a large number of authenticated and well-characterised reference samples. For each reference sample a multi-dimensional point is plotted that corresponds to the value of every component or parameter that was measured. This is analogous to plotting a point for just two parameters on an x-y graph.  Statistical pattern-recognition techniques, such as Principal Component Analysis PCA, are then used to see whether the reference samples fall into clusters, depending upon their provenance.  An example is in Figure 3[6], PCA of mass spectra from reference samples of different species of ground dried herbs.

The test sample is then measured and plotted in a similar way.  If it purported (in this example) to be oregano but the PCA plot did not fall within the reference cluster of oregano samples then suspicions would be raised (in this example, and expert microscopist might be able to confirm if the product had been diluted with e.g. olive leaves).

Generally the authentic database approach is better at confirming or otherwise the claimed origin or authenticity of a sample rather than determining the origin or authenticity of an unknown sample.  

The strength of MVA is that there is no pre-conception about what the fraudulent activity or problem might be – just that the sample is different than the reference set. If results can be plotted on a visual graph then this gives the advantage of an instinctive appreciation of how wide a difference between the test sample and the reference cluster.  Such a crude probabilistic interpretation is invaluable for prioritising resources to follow-up audit or investigation.

The limitation of all MVA approaches is the strength of the reference database.  Is it representative of all-natural variation within genuine examples, in terms of provenance, of the food in question?  It is difficult to predict the effect of seemingly minor variations on the position of an MVA data point in a pattern, particularly when the parameters being measured are uncharacterised with no cause-and-effect theories underpinning their variation.  For example, the MVA pattern of fats in beef, intended to diagnose the cattle breed, might be profoundly affected by a change in the composition of cattle feed.

Reference datasets are often built in-house by laboratories, with the risk that they do not appreciate the full nuances and variety of the genuine food on the market, and so unwittingly exclude some variations related to provenance.  The best reference datasets are constructed in collaboration with the appropriate food industry. To ensure effectiveness, reference databases can need continuous updating, particularly where the measurement principle may be affected by seasonal/yearly changes e.g. temperature and rainfall.

Reference databases are expensive to construct, a big commercial investment for any laboratory. Due to both practical and cost limitations the number of individual data points in a reference database can be limited.  There are valid technical reasons why some datasets cannot be transferred between different instruments in different laboratories, but there are also Intellectual Property protections on some reference datasets.  This can make it difficult to challenge or to gain a second opinion on test results and interpretations.  It also means that different laboratories specialise in different applications, and even different food types.  There are programmes to co-ordinate different laboratory offerings to provide virtual networks of expertise, for example the Food Authenticity Network[7] in the UK.

Laboratory vs point-of-use testing

The traditional analytical model is for samples to be sent to a laboratory, with results returned in a few days or weeks.  This is beginning to change.  There are clear advantages to the food industry in tests that can be conducted at point of use and give a real-time result.  Some such tests are now in routine use. 

One example, used in the processed fish industry, is online NMR to tell the species of frozen white block fish ingredients.   Another is the use of Near Infra Red (NIR) scanners for authenticity testing of milk coming into large dairy collection centres used by milk-powder manufacturers.  NIR and Ultra-violet / visible light (UV-Vis) are particularly suited for hand-held scanners and continuous flow applications (see Information Statement “Food Authenticity Testing: Analytical Techniques”), and have widespread use in the pharmaceutical industry for similar on-line verification testing of raw materials.  A real-time warning flag is raised if there is anything abnormal about the raw material intake. The key to this type of test is that a food company should set up this testing for their own product lines using authentic samples. The system should also be set-up to monitor how products change over time, with suitable flags to alert when something significant has changed in the manufacturing process or supply chain. This change may not always be the result of fraud, but simply due to a known swap such as the supplier or variety of ingredient. The actual samples size used by these "point-of use” methods is often small - therefore averaged data from replicate analyses, or good sample homogenisation may be required before analysis.  This is especially true for non-liquid product types.

The next predicted paradigm shift is towards tests that can be conducted by the general public at supermarket shelves or in their own homes, and the concept of Citizen Science popularised by astronomy.  Verification of “free-from” claims is a particular area of interest, using small immunoassay test kits (analogous to home pregnancy test kits) linked to smartphones to upload results onto public databases.  There is at least one kit already on the market[8] and, although as currently sold these cannot be relied on for valid results, there are major publicly-funded research projects[9] to develop and validate home allergen test kits or miniaturised NIR-scanners linked to smartphones that will pass scientific acceptance.

The majority of testing will continue to be conducted in specialist laboratories in the foreseeable future, due to the inherent capital cost of equipment, need for purchase and disposal of specialist reagents, a highly controlled environment, the need for expert interpretation, or due to simple economies of scale.  But the use of certain tests in limited applications within food production-line environments has provided a step-change in the effectiveness of fraud detection measures in recent years.

Example applications of analytical testing

Once it is known that a particular food fraud can be detected, and that testing is in routine use, fraudsters will move on to something else.  Therefore, whilst the established validated methods are needed for continued due diligence, development of new methods tends to be a rapidly moving field.  There can be a necessary compromise between speed of development, publication, offering to market, and the robustness and scope of the method validation.

Many of the established, validated and documented methods were developed under publicly-funded programmes such as the UK food authenticity programme (methods now curated online7) or the EU Food Integrity Programme.

Table 1 gives some examples of the many publications on different test methods and applications.  This is to give a flavour of what is available; it is not a comprehensive list of the thousands of scientific papers that have been published on specific test methods and applications in recent years, nor an IFST endorsement of a particular method or researcher.  It is also important to remember the role of lower-technology testing in food analysis.  Some claims (e.g. “Premium muesli, with 50% fruit”) can be verified simply by separating and weighing the ingredients, and despite the best advances of science, there is no substitute for an expert taste panel to verify the provenance of premium whiskies, wines, or olive oils.

Some of the most powerful analytical applications use a combination of different test techniques, and perform multivariate analysis on the total data set.

Table 1: Example Applications of Analytical Testing

Food

Issue

Technique

Targeted?

Ref.

Almonds (ground)

Peanut addition

ICP-OES

Untargeted

[10]

Butter

Palm/coconut oil addition

Fluorescence spectroscopy

Untargeted

[11]

Cheese

Plant oil addition

Fluorescence spectroscopy

Untargeted

[12]

Coffee

Arabica vs Robusta

NMR

Untargeted

[13]

Cooking oils

Variety substitution

FTIR

Untargeted

[14]

Fish

Catch area

DNA PCR-RAPD

Untargeted

[15]

Fish

Species substitution

MALDI-ToF-MS with proteomics

Untargeted

[16]

Fruit juice

Apple juice addition

HPLC

Targeted

[17]

Fruit (plums)

Organic production & cultivar verification

Mass spectrometry (of volatiles)

Untargeted

[18]

Grains

Authenticity

NIR & UV-Vis spectroscopy

Untargeted

[19]

Herbs (dried)

Authenticity

FTIR & mass spectrometry

Untargeted

[20]

Honey

Sugar addition

NMR

Targeted

[21]

Honey

Sugar addition

Isotope ratio MS

Targeted

[22]

Honey

Floral origin

Mass spectrometry & Isotope ratio MS & Raman & NIR spectroscopy & ICP-MS

Untargeted

[23]

Margarine

Fat profile

Raman & NIR spectroscopy

Untargeted

[24]

Meat

Species adulteration

NIR & UV-Vis spectroscopy

Untargeted

[25]

Meat

Adulteration with offal

IR spectroscopy

Untargeted

[26]

Meat

Species adulteration

DNA hybridisation probes

Targeted

[27]

Meat

Species adulteration

DNA PCR amplification

Targeted

[28]

Milk powder

Melamine

NIR spectroscopy

Targeted

[29]

Milk

Nitrogen enrichment

Mass spectrometry

Targeted

[30]

Milk

Additives for shelf life extension, dilution

IR spectroscopy

Targeted

[31]

Olive oil

Geographic origin

Mass spectrometry

Untargeted

[32]

Parmigiano Reggiano

Fatty acid profile

Mass spectrometry

Targeted

[33]

Rice

Variety mislabelling

DNA PCR-RAPD

Untargeted

[34]

Salt

Premium origin

NIR spectroscopy

Untargeted

[35]

Shellfish

Geographic origin

ICP-AES

Untargeted

[36]

Vegetables

Organic production

Isotope ratio MS

Targeted

[37]

Vinegar

Wine vinegar authenticity

Isotope ratio MS

Targeted

[38]

See IFST Information Statement “Food Authenticity Testing: Analytical Techniques” for further details on specific techniques.

 

Conclusions

There is a conceptual divide between analytical techniques traditionally used for food contaminants or nutritional parameters and those used for many food authenticity tests. Rather than measure a specific component against a fixed limit, they often rely on a probabilistic match of a result or a pattern of results against a reference database of authentic samples.

This means that the interpretation of modern authenticity test results rarely meets the burden of proof that would be required in a court of law. There is inevitable uncertainty over both the fitness of the probability match and whether the reference database is truly representative of the test sample.

Provided that these caveats are appreciated, authenticity testing has a valuable place in the supply chain assurance programmes of food businesses. Test results can be used to target and inform follow-up investigations and audits. And testing programmes are a deterrent to potential fraudsters.

The analytical techniques and references databases used for authenticity testing are rapidly evolving.  Whatever the authenticity question there is likely a research group, somewhere, working on it. Amongst the plethora of scientific publications and advertised laboratory services, it is important to differentiate between proof-of-concept studies using narrowly controlled conditions and approaches that have been applied to real-world situations.

Rather than a transactional customer-client relationship for analytical testing, laboratories are increasingly working with food industry clients to understand and tailor analytical approaches to address their specific authenticity risks and ingredient or product types. This collaboration and communication is often essential for the successful interpretation of results.

Before commissioning analyses, food companies should ensure that they;

  • understand the scope of the reference database used,
  • understand the likely outputs/information which will be obtained from a particular food authenticity test, and
  • document the next steps to be taken in case of a suspicious result.

Glossary

AIJN

The representative association of the fruit juice industry in the European Union

DNA

Used generically to describe test methods that identify based on protein sequences within nucleic acids

ELISA

Enzyme Linked ImmunoSorbent Assay

GC-MS

Gas Chromatography – Mass Spectrometry

HADH

An oxidoreductase enzyme, 3-hydroxyacyl-CoA dehydrogenase

HPLC

High Performance Liquid Chromatography

HMF

Hydroxymethyl formamide, formed when honey is overheated

LC-MS

Liquid Chromatography – Mass Spectrometry

MVA

Multivariate Analysis, a statistical treatment of data clusters

NIR

Near Infra-Red spectroscopy

NMR

Nuclear Magnetic Resonance spectroscopy

PDO

Protected Designation of Origin

UV

Ultra violet light absorbance spectroscopy

References

1. J Points, in “Horizonscan occasional articles 4: Food and feed authenticity – recent trends”, published by Fera Science Ltd, York, UK, 2016.

2. “Elliot review into the integrity and assurance of food supply networks – final report”, HM Government PB 14192, July 2014

3. “Global Standard Food Safety Issue 8”, British Retail Consortium, London, UK, 2018

4. John Spink and Douglas Moyer, Defining the Public Health Threat of Food Fraud, Journal of Food Science, 76 (2011) R157

5. MP Calloa and I Ruisanchez, “An overview of multivariate qualitative methods for food fraud analysis”, Food Control 86 (2018) 283

6. SA Haughey et al, “A comprehensive strategy to detect the fraudulent adulteration of herbs: The oregano approach”, Food Chemistry 2010 (2016) 551

7.  http://www.foodauthenticity.uk/, accessed 4 January 2018

8. https://glutentox.com/, accessed 19 December 2017

9. http://www.foodsmartphone.eu/, accessed 19 December 2017

10. M Esteki et al “Qualitative and quantitative analysis of peanut adulteration in almond powder samples using multi-element fingerprinting combined with multivariate data analysis” Food Control 82(2017) 31

11. A Dankowska et al, “Application of synchronous fluorescence spectroscopy with multivariate data analysis for determination of butter adulteration”, Int J Food Sci Tech, 49 (2014) 2628

12. A Dankowska et al, “Dectection of plant oil addition to cheese by synchronous fluorescence spectroscopy”, Dairy Sci and Tech 95 (2015) 413

13. DW Lachenmeier et al, “Rapid approach to identify the presence of Arabica and Rustica species in coffee using 1H NMR spectroscopy”, Food Chemistry 182 (2015) 178

14. MZ Durak et al, “Rapid detection of adulteration of cold pressed sesame oil adulterated with hazelnut, canola and sunflower oils using ATR-FTIR spectroscopy combined with chemometric” Food Control 82 (2017) 212

15. M Fischer et al “Applying population genetics for authentication of marine fish: the case of saithe (pollachius virens)” J Agric Food Chem 63 (2015) 802

16. A Stahl and U Schroder, “Development of a MALDI-ToF MS based protein fingerprint database of common food fish allowing fast and reliable identification of fraud and substitution”, J Agric Food Chem, 65 (2017) 7519

17. R Vanderlinde et al, “Detection of addition of apple juice in purple grape juice”, Food Control 69 (2016) 1

18. J Moreno-Rojas et al “Effect of management (organic vs conventional) on volatile profiles of six plum cultivars a chemometric approach for varietal classification and determination of potential markers” Food Chemistry 199 (2016) 479

19.  M Burns et al “Feasibility study for applying spectral imaging for wheat grain authenticity testing in pasta” Food Nutrit.Sci. 7 (2016) 355

20.  SA Haughey et al, “A comprehensive strategy to detect the fraudulent adulteration of herbs: The oregano approach”, Food Chemistry 2010 (2016) 551

21. E Jamin et al, “Fast and global authenticity screening of honey using 1H-NMR profiling”, Food Chemistry 189 (2015) 60

22. M Tosun, “Detection of adulteration in honey samples added various sugar syrups with 13C/12C isotope ratio analysis method” Food Chemistry 138 (2013) 1629

23. Z Jandric et al, “Discrimination of honeys of different floral origins by a combination of various chemical parameters”, Food Chemistry 189 (2015) 52

24. D Ucuncuoglu et al, “Rapid detection of adulteration in bakery products using Raman and near infrared spectroscopies”, Eur Food Res Technol 237 (2013) 703

25. JE Nychas et al, “Multispectral imaging: a promising method for the detection of minced beef adulteration with horsemeat” Food Control 73 (2017) 57

26. Y Hu et al, “Detection and quantification of offal content in ground beef meat using vibrational spectroscopic-based chemometric analysis” Scientific Reports 9 November 2017  DOI:10.1038/s41598-017-15389-3

27. S Rahmati eg al, “Identification of meat origin in food products: a review”, Food Control 68 (2016) 379

28. S Rahmati eg al, “Identification of meat origin in food products: a review”, Food Control 68 (2016) 379

29. PF Scholl et al, “Effects of the adulteration technique on the near-infrared detection of melamine in milk powder” J Agric Food Chem 65 (2017) 5799

30. N Frank et al “Development of a quantitative multi-compound method for the detection of 14 nitrogen-rich adulterants by LC-MS/MS in food materials” Food Add Contam A (2017) DOI:0.1080/19440049.2017.1372640

31. M Sena et al “Development and analytical validation of a screening method for simultaneous detection of five adulterants in raw milk using mid-infrared spectroscopy and PLS-DA” Food Chemistry 181 (2015) 31

32. R Gil-Solsona et al, “Metabolomic approach for extra virgin olive oil origin discrimination making use of ultra high performance liquid chromatography with quadrupole time-of-flight mass spectrometry” Food Control 70 (2016) 350

33. A Caligiani et al, “Development of a quantitative GC-MS method for the detection of cyclopropane fatty acids in cheese as new molecular markers for Parmigiano Reggiano authentication”, J Agric Food Chem, 64 (2016) 4158

34. M Arlorio et al “Chemometrical characterisation of four Italian rice varieties based on genetic and chemical analysis” J Agric Food Chem 54 (2006) 9985

35. A Rangel et al, “Fourier transform near infra-red spectroscopy application for sea salt quality evaluation” J Agric Food Chem, 59 (2011) 11109

36. L Li et al “Assessment of elemental profiling for distinguishing geographic origin of aquacultured shrimp from India, Thailand and Vietnam” Food Control 80 (2017) 162

37. S Kelly et al “Nitrogen isotope composition of organically and conventionally grown crops” J Agric Food Chem 55 (2007) 2664

38. F Camin et al “Control of wine vinegar authenticity through ɖ18O analysis” Food Control 29 (2013) 107

Institute of Food Science & Technology has authorised the publication of the following updated Information Statement entitled 'Food authenticity testing part 1: The role of analysis' dated February 2019, replacing that of November 2013.

This updated Information Statement has been prepared by John Points MIFST, peer reviewed by professional members of IFST and approved by the IFST Scientific Comittee.

The Institute takes every possible care in compiling, preparing and issuing the information contained in IFST Information Statements, but can accept no liability whatsoever in connection with them. Nothing in them should be construed as absolving anyone from complying with legal requirements. They are provided for general information and guidance and to express expert professional interpretation and opinion, on important food-related issues.