Fake Meat: What Is Fraud, and What Is Simply a Different Food

“Fake meat” is two completely different subjects wearing one name, and almost every argument about it is really an argument about which subject the other person meant. One of them is a large, thoroughly documented body of food fraud: meat sold as an animal it did not come from, meat sold with undeclared water and protein pumped into it, meat sold under a farming claim nobody verified, and meat made to look fresher than it is. The other is a shelf of honestly labelled products — plant protein shaped into a patty, a fungal protein mince, a piece of muscle grown from animal cells in a tank — that say on the front of the pack exactly what they are.

Only the first is fraud. This page is mostly about the first, because that is where the evidence is and where a reader can actually be harmed. But it starts with the second, because the confusion between them is doing real damage: it lets genuine meat fraud hide behind an argument about vegetarianism, and it lets a shopper who deliberately chose a labelled meat analogue be told they were tricked. Nobody was tricked. They read the label and bought the thing it described.

Table of Contents

  1. What Is Not Fraud, Stated Plainly
  2. What Counts as Meat, Legally
  3. Species Substitution: The Best-Evidenced Meat Fraud
  4. The 2013 European Horsemeat Incident, Accurately
  5. Seafood: The Highest Documented Substitution Rates
  6. Extension: Water, Salt, Phosphate and Borrowed Protein
  7. Mechanically Separated Meat and the Word “Meat”
  8. Provenance and Welfare Claims: What the Words Require
  9. Dates, Colour and the Appearance of Freshness
  10. Halal and Kosher Certification Fraud
  11. How Laboratories Actually Detect It
  12. Where Every One of Those Methods Fails
  13. Plant-Based Analogues: The Honest Nutrition Comparison
  14. Protein Quality, Iron and B12, Without Overclaiming
  15. Fraud Inside the Analogue Aisle
  16. Cultivated Meat: A Labelled Product, Not a Fake
  17. What a Shopper Can Actually Do
  18. The Bottom Line
  19. Research Papers
  20. Connections
  21. Featured Videos

What Is Not Fraud, Stated Plainly

A plant-based burger sold as a plant-based burger is not fake food. Neither is a mycoprotein mince, a pea-protein sausage, a seitan strip, or a cultivated chicken fillet grown from cells — provided each is labelled as what it is. These are different products, honestly described, bought on purpose by people who wanted them. Whatever you think of how they taste or how heavily they are processed, the transaction is straight.

This matters more than it sounds, for three reasons.

First, the law agrees. In October 2024 the Court of Justice of the European Union held (Case C-438/23) that a member state may not prohibit plant-based products from using customary or descriptive meat-style names where no legal name has been reserved. The judgment turned on a simple point: the consumer-information rules already forbid misleading labelling, and a clearly-marked vegetarian product carrying a familiar name is not misleading. In the United States, several states passed laws restricting meat vocabulary on plant products from 2018 onwards, and at least one was preliminarily blocked in federal court on free-speech grounds. Regulators are still tightening how these names must be qualified — the FDA issued draft guidance in early 2025 recommending that names of plant-based alternatives identify the plant source — but the direction is towards clearer naming, not towards calling the products fraudulent.

Second, mixing the two subjects protects the fraud. Every hour spent arguing about whether a veggie burger may be called a burger is an hour not spent on the pork DNA in a product sold as halal, or the 32% mislabelling rate found across a national finfish supply chain, or the undeclared water in a chicken breast. Species substitution is a criminal problem with a large peer-reviewed literature. The naming of an honestly-labelled analogue is a marketing dispute. Treating them as the same story is how the first one gets away.

Third, the analogue aisle has its own genuine fraud problem — and it runs in the opposite direction from the one people expect. What turns up in the laboratory is meat found in products sold as vegan, not the reverse. There is now a validated method for exactly that (PMID 40294441), and we cover it below. A vegan who eats undeclared pork has been defrauded just as surely as a shopper who buys horse labelled as beef.

So: no sneering in either direction on this page. If you want to eat meat, eat meat. If you want to eat a pea-protein patty, eat a pea-protein patty. What you are entitled to, in both cases, is that the label is true.

Back to Table of Contents

What Counts as Meat, Legally

You cannot describe meat fraud without first knowing what the word is legally allowed to cover, and the answer is narrower and stranger than most people assume.

The European definition, with numbers attached

In the European Union, Regulation (EU) No 1169/2011 — the food-information rules — defines “… meat” in an ingredients list as skeletal muscle with naturally included or adherent fat and connective tissue, and then caps how much of that fat and connective tissue may ride along:

Above those limits, the material is still edible and still legal — but it may no longer all be counted as “meat.” The declared meat content has to be adjusted downwards, and the ingredients list must separately name the extra fat and connective tissue. That single rule is why a sausage can list “pork 60%, pork fat, pork connective tissue” instead of “pork 80%.” The second version would be the lie.

The same regulation excludes mechanically separated meat from the definition of meat entirely, and adds three labelling duties that exist purely because each corresponds to a fraud somebody committed:

The American definition

United States rules take a different route to a similar place. Federal inspection regulations define meat, define mechanically separated species products separately, and cap the bone-solids content of mechanically separated pork and poultry — using calcium as the routine proxy measurement, because calcium is what bone leaves behind. Mechanically separated beef has been prohibited for human food since 2004. Raw poultry sold with an injected solution must carry the percentage and the solution's ingredients inside the product name, which is why a pack reads “chicken breast, contains up to 15% of a solution of water, salt and sodium phosphate” rather than simply “chicken breast.”

Both systems share a design principle worth internalising: the rules do not ban cheap ingredients, they ban silent ones. Almost nothing on this page is fraud because of what was in the product. It is fraud because of what the label did not say.

Back to Table of Contents

Species Substitution: The Best-Evidenced Meat Fraud

If you want one meat fraud that is past argument, it is species substitution — selling one animal as another. It is the best-documented, because DNA makes it checkable. A cow, a horse, a pig and a chicken carry different genomes; a laboratory that can read a short diagnostic stretch of DNA can say which animal a sample came from without caring what the packet claimed.

The workhorse technique is DNA barcoding: amplify a standard gene region — usually part of the mitochondrial cytochrome c oxidase I (COI) gene, or cytochrome b — sequence it, and match the sequence against a reference library of known species. The method was designed for taxonomy and turned out to be a near-perfect food-fraud tool, because it answers the one question a label can lie about most profitably (PMID 33530758).

Where a laboratory already knows which substitution it is hunting for, it can skip sequencing and use species-specific PCR instead: primers designed to amplify only horse, or only pork, giving a yes/no answer in a couple of hours. Field-deployable versions now exist — a PCR reaction read on a lateral-flow strip, like a pregnancy test for horse DNA (PMID 33601658).

What the surveys keep finding

Across dozens of published surveys, the same pattern repeats. Whole cuts of muscle are rarely substituted — you can see what they are. Substitution concentrates where the product is minced, blended, formed, breaded, cooked or spiced, because those processes destroy the visual and textural evidence. Ground and processed products, sausages, kebab meat, burgers, meatballs, ready meals and pet-food-adjacent categories are where undeclared species turn up.

The economics are the same everywhere: the substitute is cheaper than the declared species, the difference is invisible once minced, and the fraud is diluted enough to survive a taste test. Game-meat products are a particularly consistent problem, because the price premium over farmed pork is large and the flavour is strong enough to cover the swap. One recent method development was aimed squarely at that: a nuclear marker in the PLAG1 zinc-finger gene that can quantify how much domestic pig is in a product sold as wild boar, rather than merely detecting its presence (PMID 40596211).

That distinction — detect versus quantify — is the whole difficulty of the field, and we return to it below.

Back to Table of Contents

The 2013 European Horsemeat Incident, Accurately

This is the canonical case, and it is worth telling precisely, because both the alarmed version and the dismissive version get it wrong.

What happened. In January 2013 Irish food-safety authorities announced that beef burgers on sale in Ireland and the United Kingdom contained equine DNA. Some samples were not trace contamination: a proportion of the product was horse. Within weeks, undeclared horsemeat had been found in processed beef products — burgers, lasagne, meatballs, ready meals — in a long list of European countries. The supply chains involved ran through multiple traders and processors across several borders, which is precisely why nobody had noticed.

What the coordinated testing found. The European Commission recommended a coordinated control plan, and its results are the most useful numbers anyone has. Of roughly 7,259 samples of beef-labelled products tested for horse DNA at or above 1%, about 193 — some 2.7% — were positive. Separately, about 3,115 horse carcases were tested for the veterinary anti-inflammatory phenylbutazone, and roughly 0.5% were positive.

What it means, honestly. Two-point-seven percent is both larger and smaller than the headlines implied. It is far too high for a supply chain that is supposed to know what animal its product came from — and it is not the “most of your mince is horse” picture that circulated at the time. The fraud was overwhelmingly economic: horse was cheaper than beef, and it was laundered through enough intermediaries that the paperwork stopped meaning anything.

But it was not purely economic, and this is the part that usually gets dropped. Horses in Europe are frequently kept as sport or companion animals rather than as food animals, and are treated with veterinary drugs that are prohibited in the human food chain. Phenylbutazone is the emblem of that problem: a drug withdrawn from routine human use decades ago because of rare but serious blood disorders. The measured contamination rate was low and the exposure any individual received was small. The point is structural rather than acute — a food chain that cannot say which species it is handling also cannot say which drug regime that animal was under. A recent veterinary review sets out that residue problem in detail (PMID 36662603).

What it changed. Analytical capacity, mostly. The incident triggered a wave of method development, of which the clearest example is a monoclonal-antibody ELISA able to detect horse meat down to 1% in raw, cooked and autoclaved ground beef or pork — useful precisely because DNA is damaged by the heat treatments that antibodies survive. The same paper is candid about the method's own flaw: the antibodies cross-react with raw poultry, so samples must be heated before analysis to avoid false positives (PMID 25474205). That is a fair summary of the whole field — every method that gains one capability gives up another.

Back to Table of Contents

Seafood: The Highest Documented Substitution Rates

If meat substitution is well documented, seafood substitution is extraordinarily well documented, and the rates are much higher. Fish is the ideal fraud commodity: a fillet has no fur, no feathers and no skeleton to identify it, hundreds of species are commercially traded, common names are inconsistent between countries, and the price gap between a prestige species and a cheap look-alike can be several-fold.

Three findings give the shape of it.

Restaurant fish, sampled over four years. DNA barcoding of nine common sushi fish ordered from 26 restaurants in one American city between 2012 and 2015 found a consistent 47% mislabelling rate — 151 of 323 samples. Every single restaurant sampled sold at least one mislabelled item. Rates were not uniform across species: halibut, red snapper, yellowfin tuna and yellowtail were mislabelled the large majority of the time, while salmon and mackerel rarely were. Sixteen samples from high-end grocery stores were mislabelled at 42%, only slightly better than the restaurants (PMID 28075039).

The supply chain, not just the counter. A Canadian study barcoded 203 specimens of 12 targeted finfish species collected from importers, registered processing plants and retailers, and found an overall 32.3% mislabelling rate — with rates differing significantly between importers and retailers, meaning the substitution was happening at more than one point. It also audited the chain-of-custody paperwork and found discrepancies in 43 samples, 21.4% of the total. Seven mislabelled samples had arrived correctly labelled under United States naming rules and became mislabelled simply by crossing a border into a different naming system (PMID 31108801).

That last detail deserves emphasis, because it is the most common honest explanation of a mislabelling statistic: not all mislabelling is fraud. Some of it is two countries using the same common name for different species. A survey that counts every name mismatch as deception will overstate the crime; a survey that excuses all of them will understate it. Good studies say which they did.

It is not a victimless naming problem. Analysis of United States trade, production and mislabelling data found that substituted products are more likely to be imported than the labelled product, that roughly 60% of mislabelled apparent consumption involved exclusively wild-caught species, and that substituted products systematically came from fisheries with less healthy stocks, greater impact of fishing on other species, and less effective management than the fisheries supplying the fish named on the label (PMID 33199620). A shopper paying more for the sustainable choice was, on average, buying the less sustainable one.

The pattern generalises. Barcoding of highly-valued fish sold in Europe has traced substitutions back to stocks landed elsewhere, with consequences for the fisheries in the exporting region (PMID 34075165), and integrated barcoding of processed fish products on the Chinese market found the same concentration of substitution in processed formats (PMID 38519172).

Health consequences follow the species. Substituting a high-mercury predatory species for a low-mercury one moves a real toxicological exposure onto the plate of someone who chose otherwise, and substituting an escolar-type fish for a named white fish causes a well-known and unpleasant digestive reaction. Allergen risk moves the same way.

Back to Table of Contents

Extension: Water, Salt, Phosphate and Borrowed Protein

Species substitution is the fraud people imagine. Extension is the fraud that touches far more shoppers, because it does not require any exotic ingredient at all — only water, and something to hold it.

How it works

Muscle protein will bind a surprising amount of added water if you help it. Salt and alkaline phosphates raise the pH away from meat's isoelectric point and swell the protein filaments so they hold liquid instead of losing it; tumbling or injecting the brine distributes it. Add hydrolysed protein — often collagen, sometimes from an entirely different species — and the mixture holds still more. The result is heavier, juicier-looking meat that shrinks dramatically in the pan, because most of what leaves the pan is what somebody sold you as meat.

Every element of that process is legal when declared. A pack labelled “contains up to 15% added solution” is an honest product; some shoppers prefer it, because the brine genuinely improves the texture of lean poultry. It becomes fraud at exactly the point the declaration stops being accurate: 25% solution sold as 15%, or added water sold as nothing at all.

How it is measured

Analytically this is old, robust and cheap. Three measurements do most of the work:

Undeclared protein from a different species is caught by the same tools used for species substitution — and increasingly by peptide mass spectrometry, which can identify the species of a hydrolysed protein additive that has no intact DNA left to sequence (PMID 31498909). Gelatine, the hardest case of all, has dedicated speciation methods for the same reason (PMID 30940288).

Why it matters nutritionally

Extension is not a poisoning story; it is a value and sodium story. If a chicken breast is 20% brine, then a fifth of the protein you paid for is not there, and the sodium you did not want is. Someone managing blood pressure and counting on plain chicken as a low-sodium food can be defeated entirely by an undeclared solution. Phosphate additives raise a second, separate question for anyone with reduced kidney function, since additive phosphate is absorbed far more completely than the phosphorus naturally bound in food.

Back to Table of Contents

Mechanically Separated Meat and the Word “Meat”

After the muscle has been cut off a carcase, a good deal of edible tissue remains attached to the bone. Forcing those bones through a machine under pressure recovers it as a paste. That material is mechanically separated meat (in Europe) or mechanically separated species (in the United States), and it is not a contaminant — it is a legitimate, inspected, nutritious ingredient with its own rules.

Those rules are strict for a reason. The pressure that liberates the tissue also liberates bone, in the form of fine fragments and marrow. European law excludes mechanically separated meat from the definition of “meat” outright: it cannot be counted towards a declared meat percentage and must be named as what it is. American law caps bone solids and treats mechanically separated beef as prohibited for human food. So the fraud here is not using the material — it is counting it as muscle meat, which inflates the meat percentage on the label and lowers the cost of the product simultaneously.

Detection is genuinely awkward, and the awkwardness is instructive.

The calcium proxy. Bone is calcium phosphate, so calcium content is the routine indirect indicator of bone debris. A method paper using automatic titration with a chelator of much higher calcium affinity than EDTA — necessary because calcium phosphate dissolves poorly — measured mechanically separated poultry samples spanning 62 to 2,833 ppm calcium, and found the calcium content of commercial sausages to be highly variable and only weakly correlated with their mechanically separated meat content. Other by-products confound it further: tendon ranged from 532 to 34,539 ppm, articular cartilage 1,069 to 1,704 ppm, and skin and fat 115 to 412 ppm (PMID 37764368). In other words, a calcium result alone cannot convict anybody.

The protein approach. Because the process ruptures cells and includes marrow, mechanically recovered material carries a distinctive protein signature. Proteomic work on chicken identified haemoglobin subunits and proteins resembling myosin-binding protein C as candidate markers — haemoglobin because marrow and residual blood come along with the tissue (PMID 21555191).

The honest summary: this fraud is detectable but not cleanly quantifiable, which is why enforcement mostly relies on factory records rather than on testing the finished sausage.

Back to Table of Contents

Provenance and Welfare Claims: What the Words Require

This is the category where the gap between what a shopper hears and what the label promises is widest — and where a claim can be technically true and still function as a deception. Nothing below is an accusation against any particular producer. It is a description of what each term actually obliges someone to do.

“Grass-fed”

Most people hear: this animal lived on pasture and ate grass. The term can legitimately mean considerably less than that. In the United States the federal marketing-claim standard for grass-fed was withdrawn in 2016; since then a grass-fed claim on inspected meat is approved from documentation the applicant supplies, and while third-party certification is encouraged it is not required by the claim itself. A forage-only diet also does not by itself require continuous pasture access — harvested forage counts. So “grass-fed” and “pasture-raised” are different claims, and only the second is about where the animal stood.

The good news is that this claim is one of the few that is chemically checkable after the fact. Pasture and grain leave different stable-isotope signatures in the animal's tissue, because grasses and cereals fix carbon by different photosynthetic pathways and because feeds differ in sulphur and nitrogen sources. In a controlled study, cattle fed pasture, a barley-based concentrate, silage-then-pasture, or silage-then-pasture-with-concentrate for a year (25 animals per group) could be discriminated by isotope ratio analysis of muscle, with the carbon and sulphur ratios performing best. More striking, sequential analysis along tail hair reconstructed changes in the animal's diet across more than a year before slaughter — an archival record that no paperwork can contradict (PMID 21391592).

The nutritional differences behind the premium are real but modest, and worth stating honestly. Grass-finished beef is consistently higher in the proportion of omega-3 fatty acids and conjugated linoleic acid and lower in total fat than grain-finished beef; the absolute quantities remain small in a portion, and total-fat differences vary with breed and finishing time. Two reviews lay this out without inflation (PMID 35028571, PMID 35267281). Buy grass-fed for the fat profile, the farming system or the flavour if you like — not because a portion will supply your omega-3 needs.

“Free-range” and “pasture-raised”

For poultry, the American free-range claim requires that birds have been allowed access to the outside. It does not, by itself, specify how much space, for how long, or what the outside consists of. “Pasture-raised” has no federal definition at all in the United States and is defined only by whichever private certification scheme the producer chose to join — which is why the audited certification marks matter far more than the adjective. European egg and poultry marketing standards are more prescriptive, setting stocking densities and minimum outdoor areas, which is one reason the terms are less elastic there.

“Organic”

This one is unusually well defined, and unusually well audited. Under the United States National Organic Program, organic livestock must be raised on certified organic feed, without antibiotics or added growth hormones, with outdoor access; ruminants must graze pasture for at least 120 days of the grazing season and take at least 30% of their dry matter intake from pasture during it. It is a legally enforceable standard with annual third-party inspection. Its limits are worth knowing too: organic is a production standard, not a nutrition standard, and it is not a guarantee about slaughter conditions.

“Natural,” “hormone-free” and breed claims

“Natural” in the American system means minimally processed with no artificial ingredients. It says nothing whatsoever about how the animal was raised, which is close to the opposite of what most shoppers take it to mean. “No hormones added” on pork or poultry is true but empty, because hormones are not permitted in raising hogs or poultry in the first place — which is why the claim has to be qualified with a statement saying so. Premium breed designations and protected geographical designations carry genuine, auditable requirements, but they are audited through documents and inspections at the farm and abattoir; the meat itself will not tell you the breed unless someone runs a genetic test against a breed-informative marker panel, which almost nobody does at retail.

Country of origin

Mandatory country-of-origin labelling for beef and pork muscle cuts was repealed in the United States in December 2015 after adverse trade rulings; it still applies to lamb, goat, chicken and fish and shellfish. That gap is why voluntary origin claims on beef became contentious enough that the Department of Agriculture finalised a rule requiring a “Product of USA” claim to mean the animal was born, raised, slaughtered and processed in the United States, with compliance from the start of 2026.

Geographical origin is, like feeding regime, partly checkable in the laboratory. Multi-element and stable-isotope profiling can place beef within broad regions, because the hydrogen and oxygen in an animal's tissue track local water and the sulphur and strontium track local geology. A multi-country European study using hydrogen, carbon, nitrogen and sulphur ratios showed the approach works at country scale (PMID 36985828). The same principle underpins the verification of farming systems in other animal products — isotopes have been used to separate hen-housing systems across two countries, for example (PMID 26343509). What isotopes cannot do is identify a farm, and they only work if somebody has built and maintained a reference database for the region in question.

Back to Table of Contents

Dates, Colour and the Appearance of Freshness

Meat advertises its own age. That is inconvenient for anyone holding stock that did not sell, and it produces two frauds: changing the date, and changing the colour.

Relabelling

Re-dating unsold meat — repacking it with a fresh date, or mincing it into a product with a longer permitted life — is the simplest fraud in this entire section. It needs no chemistry, no supplier, and no accomplice outside the room. It is also the hardest to detect from the product, because nothing has been added: the only evidence is documentary, which is why enforcement here is about traceability records, weight reconciliation and unannounced inspection rather than laboratory analysis.

It is worth separating this from the ordinary confusion about date labels, which is not fraud at all. A “best before” date is a quality estimate; a “use by” date on raw meat is a safety instruction and should be obeyed. Freezing before the use-by date stops the clock legitimately. Most household meat waste comes from misreading these, not from being defrauded.

Colour

The red of fresh meat is oxymyoglobin. As it oxidises further it becomes metmyoglobin, and the meat turns brown — a change that tracks age loosely but is not itself a safety indicator. Consumers nevertheless treat red as fresh and brown as spoiled, so anything that holds the red is commercially valuable, and the line between preservation and deception runs right through the middle of this.

Carbon monoxide. A small percentage of carbon monoxide in low-oxygen case-ready packaging binds myoglobin as carboxymyoglobin, an exceptionally stable bright cherry-red. Work on beef steaks shows how effectively pre-treatment stabilises colour during vacuum-packaged storage (PMID 28259075). It is permitted in the United States and is not authorised for meat in the European Union — and the European objection is precisely the deception question: colour stability that outlasts microbiological shelf life removes the visual cue the shopper was relying on. Used within a properly managed cold chain it is not dangerous; the argument is about what the packet is allowed to imply.

Sulphites. Sulphiting agents stabilise colour, inhibit browning and suppress microbial growth, which makes them tempting in minced meat — and their use in meat is tightly restricted for that reason. A comprehensive review of sulphites in meat sets out the pattern: legislation limits them to a narrow set of categories, yet surveys repeatedly find meat preparations to be a leading source of sulphur-dioxide exposure, particularly in adults and young people. Sulphites are associated with hypersensitivity and allergic reactions, and they destroy thiamine, so an undeclared sulphited mince is both a fraud and a nutritional subtraction (PMID 33336981).

The practical lesson for a shopper is to stop using colour as your freshness test. Smell, texture, package integrity and the use-by date are all better signals than redness, and redness is the single signal a seller can most easily manufacture.

Back to Table of Contents

Halal and Kosher Certification Fraud

For millions of readers, the most consequential meat label is not a nutrition panel — it is a certification mark, and a false one is not an economic inconvenience but a violation of religious obligation.

Two features make these marks unusually exposed. First, the property being certified is invisible in the product. Halal and kosher status depends on the species, the method and conditions of slaughter, the person who performed it, and the avoidance of prohibited substances at every subsequent step. Only the first of those leaves a physical trace. A laboratory can prove that a sample contains porcine DNA; no laboratory can confirm that a lawful slaughter was correctly performed. Second, certification is issued by many independent bodies with no single global standard, so a mark's meaning depends entirely on the rigour of whoever issued it — and a counterfeit mark on a package requires no more than a printer.

What testing can do is police the species question, and that capability has advanced sharply. Digital PCR targeting multi-copy nuclear and mitochondrial genes now detects pork DNA at very low levels in complex matrices (PMID 38190283), and the same nuclear-marker work developed for wild-boar authentication applies directly here (PMID 40596211). The hardest cases are not meat at all but derived ingredients: gelatine in confectionery and capsules, enzymes, emulsifiers, stocks and rendered fats, where processing has destroyed the DNA and only peptide-level analysis can assign a species (PMID 30940288, PMID 31498909).

The structural response has been to strengthen the paperwork rather than the chemistry: standardised certification criteria, accreditation of the certifying bodies themselves, and digital traceability from abattoir to shelf. A recent review surveys where that has reached and where the integrity gaps remain (PMID 42027979). For a reader, the practical implication is the one that holds across this whole page: a mark is only as good as the auditing behind it, so a mark issued by an accredited certifier who publishes its standard and its certificate register is worth more than an unattributed symbol, however official the symbol looks.

Back to Table of Contents

How Laboratories Actually Detect It

Five families of method carry almost all of the work. Each answers a different question, and knowing which question a method answers is how you read a food-fraud headline properly.

DNA barcoding — “what species is this?”

Amplify a standard gene region, sequence it, match it against a reference library. It is open-ended: you do not have to guess the adulterant in advance, which is why it is the discovery tool of choice and why nearly every seafood survey uses it (PMID 33530758).

Species-specific and digital PCR — “is this particular species present, and how much?”

Faster, cheaper and more sensitive than sequencing, but it only finds what you asked it to look for. Digital PCR partitions the sample into thousands of tiny reactions and counts positives, giving absolute quantification without a standard curve — a large advance for low-level detection in awkward matrices (PMID 38190283). Lateral-flow readouts move a yes/no version out of the laboratory entirely (PMID 33601658).

Proteomics and peptide markers — “what species, even after cooking?”

Liquid chromatography with tandem mass spectrometry identifies short, heat-stable peptide sequences unique to a species. This is the method of choice where DNA has been destroyed — retorted, autoclaved, hydrolysed or gelatinised products — and it can be made quantitative rather than merely presence/absence. A comprehensive review covers the approach for processed meat authentication (PMID 31498909), and label-free quantification has been applied directly to working out the species composition of processed products (PMID 28763955).

Immunoassay — “a fast, cheap screen”

Antibody-based tests are inexpensive enough for routine screening at volume, and some survive severe heat treatment, which DNA does not — the horse-detection ELISA works in autoclaved product (PMID 25474205). Their weakness is cross-reactivity between related species, so a positive is a reason to run a confirmatory method, not a verdict.

Isotope ratios and elemental profiling — “where did it live and what did it eat?”

The only family that speaks to provenance and feeding regime rather than species. Carbon and nitrogen record diet, sulphur and strontium record geology, hydrogen and oxygen record local water. Applied to feeding-system verification (PMID 21391592) and to geographical origin at country scale (PMID 36985828).

And the direction of travel

The field is moving from targeted testing — look for horse, look for pork — towards non-targeted fingerprinting, in which spectroscopic or chromatographic patterns from an authentic sample set define what normal looks like, and anything statistically abnormal is flagged for investigation. The attraction is obvious: targeted methods can only catch frauds somebody already thought of, and fraudsters move faster than method-validation committees. A comprehensive overview of emerging techniques and chemometrics for animal-derived food covers this shift (PMID 36152561).

Back to Table of Contents

Where Every One of Those Methods Fails

Detection stories are usually told as triumphs. The failure modes matter more, because they define what a “clean” test result is actually worth.

Processing destroys DNA

Heat, pressure, low pH and enzymatic hydrolysis fragment DNA. Past a certain point the target sequence is simply not there to amplify, so a negative PCR on a retorted or heavily hydrolysed product means the method could not see, not the adulterant is absent. This is why the most processed products — exactly the ones where substitution is most profitable — are the hardest to test. Shorter amplicons help; they do not fix it.

DNA copy number is not mass

This is the most consequential technical trap in the field. Most rapid species tests target mitochondrial DNA because there are many copies per cell and detection is therefore sensitive. But mitochondrial copy number varies by tissue type and by species — heart and muscle are dense with mitochondria, fat and connective tissue much less so. A mitochondrial signal therefore tells you a species is present, and tells you very little reliably about how much of the product it is. Converting a mitochondrial result into a percentage is where laboratories and journalists both go wrong. The fix is single-copy nuclear markers, which is precisely the argument of the PLAG1 work (PMID 40596211) and part of the case for digital PCR (PMID 38190283).

Blends and thresholds hide small fractions

A product built from four species at 25% each defeats any test designed around a single expected adulterant, and dilution puts each component near the limit of quantification. Enforcement thresholds — the 1% figure used in the European coordinated plan is the well-known example — exist because genuine cross-contamination happens in shared equipment and cannot be treated as fraud. A deliberate 0.8% substitution is therefore invisible to the rule, even when it is visible to the instrument.

Cross-reactivity cuts both ways

Antibodies raised against one species bind relatives; primers designed for one species can amplify a close congener. Fish are the worst case, with hundreds of closely related commercial species and reference libraries that are incomplete for exactly the obscure species most likely to be used as substitutes. A barcode can only match what somebody has already sequenced and deposited.

Isotopes give ranges, not identities

Isotope work separates groups — this feeding regime from that one, this region from that one — with overlapping distributions at the edges. It cannot identify a farm, it degrades as an animal's diet and location change through its life, and it is worthless without a maintained reference database for the region in question. It is powerful corroboration and weak sole evidence.

And the deepest limit: most claims are not chemical at all

No instrument can test whether an animal had genuine outdoor access, whether it was slaughtered by the required method, whether the date on the label is the date it was packed, or whether the consignment that arrived is the consignment on the certificate. Those claims live entirely in documents, and documents are what failed in 2013. That is why traceability systems, unannounced audits and whistleblower protection do more for label honesty than any laboratory does — and why the food-fraud literature has moved towards vulnerability assessment: mapping which points in a supply chain create the opportunity, rather than waiting to test the product at the end of it.

Back to Table of Contents

Plant-Based Analogues: The Honest Nutrition Comparison

Now the other subject. These products are not fraud, so the only fair question is a nutritional one — and it deserves a straight answer rather than a slogan from either camp.

The first and most important finding in this literature is not about sodium or protein. It is about variance. A survey of 475 plant-based meat alternatives and 754 meat products on the United Kingdom market found significant differences between the categories in price, fat, saturated fat, carbohydrate, sugar, protein, fibre and energy — but emphasised the considerable variation between product categories, so that substituting like-for-like changes intakes in ways that depend heavily on which product you picked (PMID 41184290). A separate United Kingdom analysis reached the same conclusion from the other direction: plant-based products scored better on a front-of-pack nutrient profiling system than their animal counterparts on average, but with greater variability within every food group (PMID 38531295).

That kills the category-level verdict. “Plant-based meat is healthy” and “plant-based meat is ultra-processed junk” are both untestable, because the spread inside the category is wider than the gap between categories. The useful unit of judgement is the individual product's label.

Sodium: high, but check what you are comparing to

Analogues are salty. Salt does structural work in these products as well as flavour work, and there is no version of a savoury formed patty that is low in it. But the comparison that gets quoted is usually the wrong one. A study of 125 meat substitutes on the Brazilian market found that vegan products did not differ from their meat counterparts in sodium — while noting that both carry high amounts in the context of a whole meal (PMID 35769773). The honest statement is therefore: a plant-based sausage is a high-sodium processed food, and so is the sausage it replaced. Compared to a plain roast joint or a piece of grilled fish, both lose.

Where sodium genuinely does shift is in a whole-diet substitution. In an 8-week randomised controlled trial in Singapore in which 89 adults replaced habitual protein foods with either animal meats or matched plant-based analogues, the analogue group's sodium intake was significantly higher — alongside significantly higher intakes of thiamine, folate, vitamin B6, calcium, potassium, magnesium and iron (PMID 41785660). Both halves of that sentence are true and neither should be dropped.

Fibre, fat and what is actually in them

Plant-based products consistently contain more fibre than the meat products they replace — that finding is about as robust as anything in the field, and it is a genuine advantage, since fibre is the nutrient most Western diets are shortest of. Carbohydrate is higher for the same reason. Total and saturated fat differences are inconsistent: coconut oil is a common structuring fat in these products, so a plant-based patty is not automatically lower in saturated fat than beef, and some are higher.

The ingredients themselves are worth understanding rather than fearing. A formed analogue is typically a protein isolate or concentrate — pea, wheat gluten, soy, fava, potato — texturised by extrusion, bound with methylcellulose or a starch, coloured with beet or a legume-derived haem protein, structured with a solid fat, and flavoured (PMID 38672930). Mycoprotein is a different thing again: filamentous fungal biomass, naturally fibrous, and consequently high in fibre without needing much added.

The reasonable criticism is not that any of these ingredients is dangerous. It is that a product built from isolates has had the food matrix taken apart and reassembled, and matrix affects digestion, satiety and mineral absorption in ways a nutrition panel does not capture. That criticism applies with equal force to a reconstituted chicken nugget, which is why it is a criticism of ultra-processing rather than of plants.

What the intervention trials show

Two trials are worth knowing, and the second one qualifies the first.

In a randomised crossover trial, 36 generally healthy adults ate at least two servings a day of plant-based alternative products for eight weeks and animal meat for eight weeks, keeping the rest of the diet as similar as possible. During the plant phase, LDL cholesterol was lower (109.9 versus 120.7 mg/dL), weight was lower (by about 0.9 kg), and trimethylamine-N-oxide was lower overall. No cardiovascular risk factor worsened (PMID 32780794). The authors reported a significant order effect on the TMAO result — it held clearly only in participants who received the plant phase second — which is exactly the sort of caveat that gets dropped when this trial is cited, and should not be.

A separate randomised trial substituting mycoprotein for red and processed meat likewise found favourable movement in cardiovascular risk markers (PMID 37624376).

These are short trials with small samples measuring risk markers, not hard endpoints. What they support is a modest claim: substituting these products for red and processed meat does not appear to harm cardiovascular risk factors and may modestly improve some. They do not support the claim that analogues are health foods, and nobody serious has made that claim.

Back to Table of Contents

Protein Quality, Iron and B12, Without Overclaiming

This is where the honest comparison stops being flattering to plant analogues, and pretending otherwise would be exactly the kind of overclaiming this site refuses to do.

Protein quality: meat wins, and the margin is large

Protein quality has two components: whether the essential amino acids are present in the right proportions, and whether you can actually digest them. The current standard measure is DIAAS, the digestible indispensable amino acid score, where 100 means the protein fully meets requirements for the reference pattern.

A study comparing chicken breast and thigh against five commercial plant-based chicken analogues using the standardised INFOGEST digestion model found chicken ahead on every measure: protein content 26.0–33.1 g per 100 g versus 21.5–32.2; degree of protein hydrolysis 54–61% versus 20–33%; total amino acid digestibility approximately 100% versus 81–93%; and DIAAS at or above 100% for chicken versus 52–65% for the analogues (PMID 42208435). The authors attribute the gap to plant protein structure, processing-induced aggregation, and fibre and other matrix components restricting enzyme access.

That is a real and substantial difference, and a label claiming “same protein as beef” on a gram-for-gram basis is telling you about quantity while staying silent about quality.

Two things soften it without erasing it. First, formulation matters enormously: a pig-model study of true ileal digestibility found that an air-classified quinoa flour and pea protein blend reached a DIAAS of 89% against the older-children-and-adults reference pattern after high-moisture extrusion — up from 80% for the unextruded flour mix — with sulphur-containing amino acids the first limiting group (PMID 41269145). Extrusion improved digestibility rather than damaging it, and complementary blending closed much of the amino-acid gap. Second, nobody eats one protein. In a mixed diet that includes legumes, whole grains such as brown rice, nuts, seeds and eggs or dairy, the limiting amino acid of any single item stops being the binding constraint. The traditional pairing of beans with brown rice exists because it works.

Iron: the picture is more interesting than either side admits

Meat's iron advantage is real and mechanistic. Haem iron is absorbed by a dedicated pathway at roughly 15–35% efficiency regardless of what else is on the plate, whereas non-haem iron from plants runs at roughly 2–20% and is strongly suppressed by phytate and polyphenols. Meat additionally carries a “meat factor” that enhances non-haem absorption from the rest of the meal.

Then the measurements complicate it. Analysis of plant-based burgers against a beef burger found beef significantly higher in total and bioaccessible iron and zinc — but when absorption was measured in a cell model, bioavailable iron from most plant-based burgers was comparable to beef. Beef stayed clearly ahead for bioavailable zinc, matched only by a mycoprotein product; the plant-based burgers were superior sources of calcium, copper, magnesium and manganese. The authors stress that the quantity of absorbable iron varies dramatically between meat alternatives (PMID 37375636).

And a randomised controlled trial adds a further qualification. Fifty-two iron-deficient women of reproductive age took a 32 mg iron supplement daily for eight weeks with a lunch containing either beef or a commercial plant-based patty. Ferritin, transferrin saturation, soluble transferrin receptor, body iron stores and haemoglobin all improved — with no difference between the two meals. The authors concluded that the meat factor does not contribute substantially in this context (PMID 40633767). The trial was funded by a cattle-industry trade body, which makes a null result that does not favour meat harder to dismiss, not easier.

The defensible position: haem iron is genuinely better absorbed, this matters most for people who are already deficient and are not supplementing, and it can be overridden by supplementation or by a well-constructed meal. Anyone with low ferritin should be measuring it rather than reasoning about it — see ferritin testing.

Vitamin B12: fortification, and the gap it leaves

B12 is made by bacteria, not by plants. An unfortified plant-based product contains essentially none, and this is the one nutrient where a straight swap can produce a genuine deficiency over time — slowly, because the liver holds years of reserve, and with neurological consequences that do not fully reverse.

Manufacturers fortify, and B12 is the single most commonly added nutrient in this category — but United Kingdom market data found it present in only about 15% of all plant-based products surveyed (PMID 38531295). It is a labelled ingredient when it is there, and it is not there most of the time. Where products are fortified the effect is measurable: in the 8-week Singapore trial, plasma B12, folate and selenium were all significantly higher in the plant-based analogue group than in the animal-meat group at week 8 (PMID 41785660).

So the practical rule is not “analogues cause B12 deficiency.” It is: fortification is a per-product fact, so read the panel, and if you are replacing most animal foods, take B12 rather than hoping the patty had it. A review of nutrient equivalence across plant-based and cultured products reaches the same conclusion for the wider set of micronutrients — iron, zinc, B12, long-chain omega-3s and creatine are where the gaps sit, and bioavailability, not just content, is the thing to look at (PMID 41470805).

Back to Table of Contents

Fraud Inside the Analogue Aisle

Having defended honestly-labelled analogues, the section they do not get to skip.

The documented fraud in this category runs the direction almost nobody predicts: meat turning up in products sold as vegan. The mechanism is mundane rather than sinister — analogues are frequently produced in facilities that also process meat, on shared lines, and cross-contamination is a real and continuous risk. But to a reader who avoids meat on religious or ethical grounds, the distinction between contamination and adulteration is small comfort; the product still was not what the label said.

Researchers at a national reference centre for authentic food developed a targeted proteomics method precisely for this. Using cross-species marker peptides and LC-MS/MS with calibrations from 0.1% to 5.0% meat by weight, they spiked vegan sausages and burger patties with pork, chicken or beef and identified four peptide markers with recovery rates of 80–120% — achieving limits of quantification below the 0.1% threshold commonly applied for unintended meat presence (PMID 40294441). A capability that did not exist a decade ago now allows a vegan claim to be audited to the same standard as a species claim on meat.

Two other honest-labelling problems belong here:

Back to Table of Contents

Cultivated Meat: A Labelled Product, Not a Fake

Cultivated meat — also called cell-cultured or cell-based meat — is animal muscle and fat cells grown in a bioreactor rather than in an animal. It is biologically meat. It is also, in every jurisdiction that has approved it, required to be labelled as what it is, which places it outside this section's definition of fake food entirely.

The regulatory position, as of this writing: Singapore was first to approve a cultivated chicken product for sale, at the end of 2020. In the United States the FDA issued “no questions” letters on the safety of cultivated chicken to two producers in 2022 and 2023, with the Department of Agriculture handling inspection and label approval for the finished product — a split jurisdiction that reflects how the American system divides cell-culture safety from meat labelling. Israel approved a cultivated beef product in early 2024. Running in the opposite direction, several American states have banned or placed moratoria on sale, beginning with Florida and Alabama in 2024 and continuing through 2025 — legislation driven by agricultural politics rather than by any safety finding.

A scoping review covering 35 studies and 45 regulatory or guidance documents found the production protocols themselves well established, and set out the hazard categories regulators are actually assessing: genetic stability of the cell lines, microbiological hazards associated with those lines, exposure to substances used in production, toxicity and allergenicity of the product or its components, post-harvest contamination, chemical residues, and nutritional risk. Its central finding is a gap rather than a danger — no standardised testing approach yet exists, so assessment currently rests on hazard analysis and good practice applied case by case (PMID 39796419).

On nutrition, the honest answer is that there is very little independent data, because there are very few products and almost no independent compositional analyses. A cultivated product's micronutrient content depends on the growth medium and on whether fat cells were co-cultured, so it is a formulation choice rather than a property of the technology — and iron, B12 and long-chain omega-3 content cannot be assumed to match conventional meat (PMID 41470805). Anyone claiming to know how these products compare nutritionally is ahead of the evidence in both directions.

What would make cultivated meat a fake-food story is a specific and foreseeable event: a cultivated product sold as conventionally farmed, or a conventional product sold as cultivated. Neither has been documented. The species-identification methods on this page would not catch the first, since the DNA is identical — that would need cell-line or growth-medium markers, which is an open analytical problem worth watching.

Back to Table of Contents

What a Shopper Can Actually Do

Most advice about spotting fake meat is worthless, because the frauds that matter are invisible by design. Here is what actually shifts the odds, ordered by how much difference it makes.

Buy meat in a form that still looks like an animal

This is the single most effective thing on the list, and it is not a trick — it is the mechanism. Substitution, extension and mechanically separated meat all require the product to be minced, formed, breaded, cured or cooked. A whole muscle cut, a bone-in joint, a whole bird, a whole fish with head and skin on: none of these can be a different species, and none can hide 20% brine. Every step of processing between the animal and the pack is a step where something could be added or swapped without you being able to see it.

Read the meat percentage and the ingredients order

On a European label, a declared meat percentage is a regulated number and separately-listed “fat” or “connective tissue” entries tell you the raw material exceeded the limits for counting as meat. “Added water,” “formed meat” and any protein of a different animal origin all have to be there if they apply. On an American label, look for the solution statement inside the product name. Every one of those phrases is a manufacturer telling you the truth because the law made them — and reading them costs nothing.

Prefer a shorter chain, and be sceptical of a long one

The 2013 incident happened in a supply chain long enough that nobody along it could see both ends. A butcher who names the farm, a fishmonger who can say which boat and which sea, a producer selling direct: none of these is a guarantee, but each replaces a document you cannot check with a person who can be asked. The corresponding warning sign is a premium claim attached to an anonymous chain — “wild” fish at a commodity price with no species name, or a prestige species in a country that lands very little of it.

Treat a certification mark as a starting point, not an answer

Ask which body issued it, whether that body publishes its standard, whether it is accredited, and whether the certificate can be looked up. Marks that survive all four questions are worth paying for. A symbol with no issuer named on the pack is worth nothing at all, and this applies equally to welfare marks, organic marks and religious certification.

Stop using colour as your freshness test

Redness is the one signal a seller can manufacture directly. Use the date, the smell, the texture and the integrity of the pack instead. Brown meat inside a vacuum pack is normal chemistry, not spoilage.

Match the product to the job

If the meal is built around the protein, buy a whole cut and cook it. If you want convenience, buy the processed product knowingly and read the sodium. And if you are replacing meat regularly, build the plate around unprocessed ingredients first — lentils, beans and chickpeas, brown rice and other whole grains, nuts and seeds, eggs, oily fish, mushrooms and plenty of vegetables — with a formed analogue as an occasional convenience rather than the foundation. That advice is not a verdict on analogues. It is the same advice we would give about sausages.

If you are replacing meat substantially, cover the four nutrients that actually move

Vitamin B12 (supplement it; do not rely on fortification you have not checked), iron (measure ferritin rather than guessing), zinc, and long-chain omega-3 fats. Those four are where the arithmetic genuinely changes. Protein quantity is rarely the problem in a Western diet; protein quality is worth a thought if your protein comes from a narrow set of sources.

Back to Table of Contents

The Bottom Line

An honestly labelled plant-based or cultivated product is not fake food. It is a different food, sold as itself, to somebody who wanted it. Arguing about whether it may be called a burger is a distraction from the frauds that are actually happening.

The real meat frauds are four. Species substitution, concentrated wherever the product has been minced or processed and running at documented rates of 30–47% in some seafood categories. Extension with undeclared water, salt, phosphate and borrowed protein. Provenance and welfare claims that promise less than shoppers hear, and sometimes less than they say. And the manufacture of apparent freshness through relabelling and colour.

Detection is real but partial. DNA answers “what species” brilliantly until processing destroys it, and answers “how much” badly unless a nuclear marker is used. Proteomics survives cooking. Isotopes reach feeding regime and broad region but never a farm. And nothing in any laboratory can verify outdoor access, slaughter method, or the date on the pack — those live in documents, which is exactly where the 2013 failure occurred.

The nutrition comparison does not go one way. Meat is clearly better for protein quality (DIAAS at or above 100% against 52–65% for some analogue categories), for haem iron, and for zinc bioavailability. Analogues are clearly better for fibre, and frequently for calcium, copper, magnesium and manganese; substituting them for red and processed meat has not harmed cardiovascular risk factors in the trials run so far and has modestly improved some. Both are high in sodium when they are processed. The variance within each category is larger than the gap between them, so read the individual product.

And one rule covers the whole page. Nothing here is fraud because of what was in the food. It is fraud because of what the label did not say. Buy things whose form you can see, read the numbers the law forced onto the pack, and ask who audits any claim you are paying extra for.

Back to Table of Contents

Research Papers

Species substitution and meat authentication

  1. Hsieh YH, Ofori JA. Detection of horse meat contamination in raw and heat-processed meat products. Journal of Agricultural and Food Chemistry. 2014;62(52):12536–12544. PMID 25474205 — a monoclonal-antibody ELISA detecting 1% horse meat in raw, cooked and autoclaved product, with its own cross-reactivity caveat stated openly.
  2. Chen Y, Wang Y, Xiao M, Wei S, Yang H, Yin R. Polymerase chain reaction with lateral flow sensor assay for the identification of horse meat in raw and processed meat products. Food Chemistry. 2021;345:128840. PMID 33601658 — species-specific PCR read on a paper strip, outside the laboratory.
  3. Weber K, Kearley ME, Marini AM, Pressman P, Hayes AW. A review of horses sent to slaughter for human consumption: impact of horsemeat consumption, residual banned drugs, and public health risks. American Journal of Veterinary Research. 2023;84(3). PMID 36662603 — why an undeclared species is also an undeclared veterinary-drug history.
  4. Nehal N, Choudhary B, Nagpure A, Gupta RK. DNA barcoding: a modern age tool for detection of adulteration in food. Critical Reviews in Biotechnology. 2021;41(5):767–791. PMID 33530758 — the method that turned food fraud into something measurable.
  5. Adenuga BM, Spychaj A, Montowska M. A new nuclear marker for quantitative analysis of wild boar and domestic pig meat in game meat products using PLAG1 zinc finger gene. Scientific Reports. 2025;15(1):20454. PMID 40596211 — the shift from detecting a species to quantifying how much of it is there.
  6. Mokhtar NFK, Shun YQ, Raja Nhari RMH, et al. Nanoplate-based digital PCR for highly sensitive pork DNA detection targeting multi-copy nuclear and mitochondrial genes. Food Additives & Contaminants: Part A. 2024;41(2):120–133. PMID 38190283 — absolute quantification without a standard curve.
  7. Riaz MN, Irshad F, Haider MB. Artificial intelligence in religious certification: a comprehensive review of halal digitalization, compliance integrity, and future directions. International Journal of Food Science. 2026;2026:3011089. PMID 42027979 — where certification integrity currently stands, and where it does not.

Seafood mislabelling

  1. Willette DA, Simmonds SE, Cheng SH, et al. Using DNA barcoding to track seafood mislabeling in Los Angeles restaurants. Conservation Biology. 2017;31(5):1076–1085. PMID 28075039 — 47% mislabelling over four years; every restaurant sampled had at least one case.
  2. Shehata HR, Bourque D, Steinke D, Chen S, Hanner R. Survey of mislabelling across finfish supply chain reveals mislabelling both outside and within Canada. Food Research International. 2019;121:723–729. PMID 31108801 — sampling at several points in the chain, which is how you find out where it happens.
  3. Kroetz K, Luque GM, Gephart JA, et al. Consequences of seafood mislabeling for marine populations and fisheries management. Proceedings of the National Academy of Sciences of the United States of America. 2020;117(48):30318–30323. PMID 33199620 — substituted products come from less healthy and less well-managed fisheries.
  4. Blanco-Fernandez C, Ardura A, Masiá P, et al. Fraud in highly appreciated fish detected from DNA in Europe may undermine the Development Goal of sustainable fishing in Africa. Scientific Reports. 2021;11(1):11423. PMID 34075165 — substitution traced back to the stocks that supplied it.
  5. Zhao S, Zhang H, Zhao Z, et al. Integrated DNA barcoding methods to identify species in the processed fish products from Chinese market. Food Research International. 2024;182:114140. PMID 38519172 — the same concentration of fraud in processed formats, in another market.

Detection methods and their limits

  1. Stachniuk A, Sumara A, Montowska M, Fornal E. Liquid chromatography-mass spectrometry bottom-up proteomic methods in animal species analysis of processed meat for food authentication and the detection of adulterations. Mass Spectrometry Reviews. 2021;40(1):3–30. PMID 31498909 — the reference review for species identification after DNA has been destroyed.
  2. Montowska M, Fornal E. Label-free quantification of meat proteins for evaluation of species composition of processed meat products. Food Chemistry. 2017;237:1092–1100. PMID 28763955 — making proteomics quantitative rather than presence/absence.
  3. Stahl-Zeng J, Sage A, Taylor P, Netto JD, Zhang T. Advances in LC-MS/MS methods for allergen testing, meat speciation, and gelatin speciation. Journal of AOAC International. 2019;102(5):1309–1315. PMID 30940288 — gelatine, the hardest ingredient in the whole authentication problem.
  4. Ye H, Yang J, Xiao G, et al. A comprehensive overview of emerging techniques and chemometrics for authenticity and traceability of animal-derived food. Food Chemistry. 2023;402:134216. PMID 36152561 — the move from targeted tests to non-targeted fingerprinting.
  5. Surowiec I, Koistinen KM, Fraser PD, Bramley PM. Proteomic approach for the detection of chicken mechanically recovered meat. Meat Science. 2011;89(2):233–237. PMID 21555191 — haemoglobin subunits as a marker of marrow coming along with the tissue.
  6. Shyichuk A, Kowalska M, Shyychuk I, Lamkiewicz J, Ziółkowska D. Determination of calcium in meat products by automatic titration with 1,2-diaminocyclohexane-N,N,N′,N′-tetraacetic acid. Molecules. 2023;28(18):6592. PMID 37764368 — and the honest finding that calcium correlates only weakly with mechanically separated meat content.

Provenance, feeding regime and freshness

  1. Osorio MT, Moloney AP, Schmidt O, Monahan FJ. Beef authentication and retrospective dietary verification using stable isotope ratio analysis of bovine muscle and tail hair. Journal of Agricultural and Food Chemistry. 2011;59(7):3295–3305. PMID 21391592 — tail hair as an archival record of what the animal ate for over a year.
  2. Bontempo L, Perini M, Pianezze S, et al. Characterization of beef coming from different European countries through stable isotope (H, C, N, and S) ratio analysis. Molecules. 2023;28(6):2856. PMID 36985828 — geographical origin at country scale, and no finer.
  3. Rogers KM, van Ruth S, Alewijn M, Philips A, Rogers P. Verification of egg farming systems from The Netherlands and New Zealand using stable isotopes. Journal of Agricultural and Food Chemistry. 2015;63(38):8372–8380. PMID 26343509 — the same isotope logic applied to a farming-system claim.
  4. Nogoy KMC, Sun B, Shin S, et al. Fatty acid composition of grain- and grass-fed beef and their nutritional value and health implication. Food Science of Animal Resources. 2022;42(1):18–33. PMID 35028571 — the real size of the grass-fed fatty-acid difference.
  5. Davis H, Magistrali A, Butler G, Stergiadis S. Nutritional benefits from fatty acids in organic and grass-fed beef. Foods. 2022;11(5):646. PMID 35267281 — a second review reaching a similarly measured conclusion.
  6. Van Rooyen LA, Allen P, Crawley SM, O’Connor DI. The effect of carbon monoxide pretreatment exposure time on the colour stability and quality attributes of vacuum packaged beef steaks. Meat Science. 2017;129:74–80. PMID 28259075 — how thoroughly colour can be decoupled from age.
  7. D’Amore T, Di Taranto A, Berardi G, et al. Sulfites in meat: occurrence, activity, toxicity, regulation, and detection. A comprehensive review. Comprehensive Reviews in Food Science and Food Safety. 2020;19(5):2701–2720. PMID 33336981 — restricted in meat, and still a leading source of sulphur-dioxide exposure.

Plant-based analogues and cultivated meat: the nutrition evidence

  1. Gouela M, Stergiadis S, Clegg ME. The nutritional composition and impact on UK dietary intakes of meat and plant-based meat alternatives. npj Science of Food. 2025;9(1):217. PMID 41184290 — 475 alternatives against 754 meat products, and the finding that variation within categories dominates.
  2. Zhang L, Langlois E, Williams K, et al. A comparative analysis of nutritional quality, amino acid profile, and nutritional supplementations in plant-based products and their animal-based counterparts in the UK. Food Chemistry. 2024;448:139059. PMID 38531295 — higher fibre, generally lower protein, and B12 fortification in only about 15% of products.
  3. Romão B, Botelho RBA, Nakano EY, et al. Are vegan alternatives to meat products healthy? A study on nutrients and main ingredients of products commercialized in Brazil. Frontiers in Public Health. 2022;10:900598. PMID 35769773 — 125 products; sodium high, but not higher than the meat products they replace.
  4. Vila-Clarà G, Vila-Martí A, Vergés-Canet L, Torres-Moreno M. Exploring the role and functionality of ingredients in plant-based meat analogue burgers: a comprehensive review. Foods. 2024;13(8):1258. PMID 38672930 — what each ingredient in a formed analogue is actually doing.
  5. Alotaibi M, Muleya M, Salter A, Hoad C, Eldeghaidy S. Comparison of the protein quality of chicken and plant-based chicken analogues using the INFOGEST in-vitro digestion system. Food Chemistry. 2026;520:149734. PMID 42208435 — DIAAS at or above 100% for chicken against 52–65% for the analogues.
  6. Cui H, Hodgkinson SM, Stroebinger N, et al. Effect of high moisture extrusion on true ileal amino acid digestibility and DIAAS of quinoa-pea protein-based meat analogues. Food & Function. 2025;16(24):9422–9432. PMID 41269145 — extrusion improved digestibility; complementary blending closed much of the amino-acid gap.
  7. Latunde-Dada GO, Kajarabille N, Rose S, et al. Content and availability of minerals in plant-based burgers compared with a meat burger. Nutrients. 2023;15(12):2732. PMID 37375636 — beef ahead on iron and zinc content, but bioavailable iron often comparable.
  8. Hennigar SR, Miller KM, Murphy RD, et al. Effects of consuming an iron supplement with a meal containing animal or plant-based meat on indicators of iron status and anemia in women of reproductive age with iron deficiency: a randomized, controlled study. The American Journal of Clinical Nutrition. 2025;122(3):859–865. PMID 40633767 — no difference in iron status between the two meals over eight weeks.
  9. Fu AS, Osman F, Cameron-Smith D, et al. Micronutrient intake and status of adults consuming plant-based meat analogues or animal-based meats as primary protein source: an 8-week randomized controlled trial. Clinical Nutrition. 2026;59:106610. PMID 41785660 — higher sodium on the analogue diet, and higher plasma B12, folate and selenium.
  10. Crimarco A, Springfield S, Petlura C, et al. A randomized crossover trial on the effect of plant-based compared with animal-based meat on trimethylamine-N-oxide and cardiovascular disease risk factors in generally healthy adults: Study With Appetizing Plantfood-Meat Eating Alternative Trial (SWAP-MEAT). The American Journal of Clinical Nutrition. 2020;112(5):1188–1199. PMID 32780794 — lower LDL and weight on the plant phase, with an order effect on the TMAO result the authors report openly.
  11. Farsi DN, Gallegos JL, Finnigan TJA, Cheung W, Munoz JM, Commane DM. The effects of substituting red and processed meat for mycoprotein on biomarkers of cardiovascular risk in healthy volunteers: an analysis of secondary endpoints from Mycomeat. European Journal of Nutrition. 2023;62(8):3349–3359. PMID 37624376 — a second substitution trial, in a different protein source.
  12. Häfner L, Jira W, Kranz B, Haase I, Bolumar T, Brockmeyer J. A rapid and robust targeted proteomics method for the quantitation of cross-contaminations and adulterations with meat in vegan and vegetarian meat analogues. Journal of Agricultural and Food Chemistry. 2025;73(18):11410–11421. PMID 40294441 — the fraud that runs the other way, quantified below 0.1% by weight.
  13. Demarquoy J. Nutrient equivalence of plant-based and cultured meat: gaps, bioavailability, and health perspectives. Nutrients. 2025;17(24):3860. PMID 41470805 — where the micronutrient gaps sit once bioavailability is taken into account.
  14. Zandonadi RP, Ramos MC, Elias FTS, Guimarães NS. Global insights into cultured meat: uncovering production processes, potential hazards, regulatory frameworks, and key challenges — a scoping review. Foods. 2025;14(1):129. PMID 39796419 — the hazard categories regulators are assessing, and the absence of a standardised testing approach.

Standards, regulations and other sources

Back to Table of Contents

Connections

Back to Table of Contents