Substantial Equivalence, Explained

[Please consider supporting GMO Building Blocks with an  ongoing contribution of $1, $2, $3, $5 or $10 a month on Patreon. Or make a one time donation via PayPal. ]

Marc BrazeauMarc Brazeau | Editor | Food and Farm Discussion Lab | @eatcookwrite

If there is one concept that drives much of divide in the GMO debate, it’s substantial equivalence. Having different understandings or misunderstandings of the concept leads to rancor, distrust and talking past each other. Proponents see it as a common sense way of determining if heightened regulatory scrutiny for a new product is warranted or unnecessary. Critics of genetic engineering often see it as some sort of trick — a sleight of hand or legalistic loophole. They ask how something can be so novel that on one hand, it merits the legal protection of a patent monopoly, and then on the other hand, the FDA can declare it to be substantially equivalent to its parent variety or breed. That suspicious and often paranoid take is nearly always based on misunderstandings of the concept of substantial equivalence and how patents work.

As long as we’re all getting on the same page …

Let’s clear up another major misconception. That is: What is the scientific consensus on the safety of GMOs? Many who are distrustful of genetic engineering or are outright anti-GMO misunderstand what the consensus is, and thus find it patently wrong on its face. And they’d be right, if the consensus was what they think it is.

More often than not, critics of the technology portray the scientific consensus as: GMOs are safe. Indeed, they can be forgiven (at least those who aren’t old pros) because proponents of the technology too often portray it in the same way.

But that isn’t just a misconception, it’s a category error. The scientific consensus is not about GMOs. It’s about genetic engineering. It’s about the process, not the products. And it’s not about “safety” it’s about relative risk.

The scientific consensus on genetic engineering is: Genetic engineering is not substantially more risky than more traditional breeding methods, and that all breeding methods carry some small risks of harmful unintended consequences. Indeed, many scientists see the relative precision of moving one or two genes to improve a plant or animal as less likely to introduce unintended consequences than traditional breeding methods. In fact, many breeding projects using traditional methods are seeking transformations that are as ambitious, if not more so, than genetically engineered traits on the market. Either way, it’s nearly unanimous among scientists in the relevant fields, that all common breeding techniques present very low risk of harmful unintended consequences, especially changes that might escape detection prior to commercialization. No breeding method is considered safe, only that the relative risk is very small. Science doesn’t deal in “safe”, only in relative risk, though the somewhat awkward language of low relative risk can get translated into “safe” for a lay readership.

With the greater regulatory scrutiny that biotech products get, a reasonable case can be made that they come with less risks for consumers, though in either case the risks are miniscule.

But there is an understanding that, just as traditional breeding methods have unintentionally raised the levels of natural toxins in crops in the past, biotech breeding could introduce or trigger new toxins or allergens. In the 1960s, celery breeders developed a variety of celery that produced higher levels of psoralens, a natural insecticide. The celery had to be pulled off the market when it turned out to give farm workers rashes due to phytophotodermatitis. Also in the 1960’s hoping for a better potato chip, breeders produced a potato the produced more of the neurotoxin solanine so that it wasn’t just toxic to insects, but produced a dose level that made humans sick. Thus there is no scientific consensus that traditional breeding methods are safe or that all products of traditional breeding are safe, only that the breeding methods are low risk, and products we regularly use are generally recognized as safe.

The GMOInside characterization of the FDA’s position in the meme above on substantial equivalence and GMOs isn’t quite correct. The FDA did not rule that all GMOs are by definition substantially equivalent and thus do not merit special labels. The rule is that new crops that are found to be substantially equivalent do not merit special labels. But the law remains that if a new crop were found to be safe for human consumption, but NOT substantially equivalent, then a label might be warranted. If Golden Rice were ever to commercialized in the US, it could very well require a label because beta-carotene is a nutrient not normally found in rice. There are lots of potential modifications that we could imagine that are safe and useful but change the composition of the plant or animal enough to put it outside of normal variation.

How is substantial equivalency used by regulators?

It should be noted that, though substantial equivalence is associated with biotech regulation, it’s a regulatory concept used by food safety agencies for analyzing a broad range of new products.

Food safety regulators in the FDA and around the world use substantial equivalency to assess how much scrutiny a new food product needs before being commercialized. If a new food is judged to be substantially equivalent to foods that are considered safe either through previous testing or, more often, through a long history of use, then they do not need to require animal feeding studies or other long and expensive demonstrations of safety (technically: a minimal level of relative risk). If a novel food product is judged to be substantially equivalent to its parent or within its breed or variety, then it can assumed to be as safe as the parent. If not, it is evaluated as a new food additive and faces further food safety scrutiny.

What does it mean to be substantially equivalent?

A judgement of substantial equivalency starts with a compositional analysis of relevant comparators. I’m going to lay out one simple example and try to get some broad conceptual points across, for a more nuts and bolts look at how compositional analysis is carried out in the context of a substantial equivalence assessment, I recommend Anastasia Bodnar’s brief essay on the matter, written by someone who has does compositional analysis. It details tools and techniques, as well as the limitations and challenges of getting a sound analysis. Please, read the whole thing, but I will quote a bit.

If there is a change that doesn’t fall within the natural variation for that species, especially if there isn’t an obvious scientific explanation for the change, then more testing needs to be done to determine safety with regard to environment and human health.

What substantial equivalence does not do is give license to make assumptions. The process of genetic engineering does have the potential to cause unintended changes in the resulting organism. That’s why a comparative assessment needs to be conducted before a plant, animal or microbe that has been genetically engineered can be deemed substantially equivalent to a non-genetically engineered but genetically similar organism.

So, if we wanted to assess a new fungus-resistant wheat, we would do compositional analysis of a wide range of samples of the parent variety, and perhaps more within the full range of commercial varieties. We’d map the range of protein and starch content, the range of metabolites and micro-nutrients produced, etc. Then we’d have a baseline of ranges for the component parts of wheat to assess substantial equivalence.

If we were looking at protein content, soft wheat usually ranges from 8-11 percent protein and hard red winter is usually between 10-13 percent. At first glance, that seems like a fairly narrow range, but higher protein wheat has 30 percent more protein than low protein wheat. And that can be within exact same variety, genetically identical, but grown under different conditions. So if a new biotech variety of winter wheat that produced either more or less protein than its parent, but still landed within the 10-13 percent range it would be substantially equivalent regarding protein content. In some cases we might zoom out to the full range of protein content of commercial wheat to assess equivalency. Within the full range, from 8-13 percent, the higher protein wheat has a full 60 percent more protein than low protein wheat. That would give a very wide range in terms of what is considered normal protein content in wheat.

Then you repeat that assessment for all the relevant properties – i.e. look into the ranges for vitamin and mineral content through the whole range of properties that this kind of compositional analysis covers. You do this until you get a full sense of whether this new wheat is wheat within the range of properties that we expect within our current supply of commercial wheat.

A 2013 paper by Herman and Price looked at 20 years of research relating to compositional changes in biotech products and found no evidence of unintended consequences in the literature. The body of research comparing the potential for unintended consequences from biotech and selective breeding due to insertional effects, copy number variation (CNV), presence/absence variation (PAV), and transcriptomic effects can be found here.

Can you pick the real tomato out of a line up?

Let’s run through some simple examples so that we lay persons can fully internalize the wide range of properties we take for granted, but then get elided when someone balks at the concept of substantial equivalence with regards to novel crops.

Tomatoes are a great example of the wide, wide, wide range of properties, of compositional variation in what we consider a tomato.

Cherry tomato mix – Wikicommons

All those greens, reds, yellows, oranges, and purples represent very different levels of various nutrients, and who knows what else? (somebody knows, probably Harry Klee and Kevin Folta at the University of Florida know but you get the point)

Scroll through and browse the tomato offerings at Johnny’s Seeds, an organic seed company that caters to gardeners and growers for the local and farm to table market. You’ll find tomatoes of all shapes and sizes bred for very distinct properties:

• Improved Indigo type with striking dark-blue anthocyanin coloration
• Late blight-resistant field cherry
• Sunny orange fruits with full flavor, meaty interior with few seeds
• Small, deep-red cherry resists late blight, cracking, and rot

In addition to the different sizes, shapes, and colors, you find tomatoes bred for cold tolerance, crack resistance, disease resistance, ease of harvest, harvest timing, and on and on. The full range of diseases they’ve bred for resistance on offer consists of alternaria blight, alternaria stem canker, bacterial speck, fusarium crown and root rot, fusarium wilt, gray leaf spot, late blight, leaf molds, powdery mildew, tobacco mosaic virus, tomato apex necrosis virus, tomato mosaic virus, tomato spotted wilt virus, tomato yellow leaf curl virus, verticillium wilt, as well as resistance to nematodes.

Heirloom tomatoes | Photo by Denise Krebs | Flickr CC license

There are two points here. The first is to hammer home just how much range in properties and composition there can be when we assess what it means for a new variety or newly traited crop to be substantially equivalent within its species. The range of sugars and proteins and flavoids and carotenoids and blight resistance and metabolites and polyphenols and gene expression among what we all recognize as tomatoes is enormous. You’d have to change the nutritional profile of a tomato quite a bit to nudge it outside of normal observable ranges. Likewise, it is possible for an organic breeder trying to increase nematode resistance to unintentionally increase the production of potentially toxic protein outside of normal ranges. It’s unlikely but not impossible, or even unheard of. The FDA would not consider that tomato to be substantially equivalent, even though bred by traditional means.

Other things to consider are the useful and idiosyncratic traits presented here that we understand as safe. If the proteins expressed in fusarium wilt resistant tomatoes are generally considered safe in tomatoes, if we were to transfer them via biotech methods to basil or bananas, is there reason to believe we introduced any more risk than in using traditional methods to another variety of tomato? The FDA would consider that as a food additive. They would then want a compositional analysis showing that the introduction of the gene for wilt resistance hadn’t resulted in other genetic changes that put the basil outside of the range of substantial equivalence.

Likewise the toxins produced by nematode resistant tomatoes to protect against nematodes might be useful in a cucumber. If the gene or genes for nematode resistance are transferred to a cucumber plant, if the compound that’s toxic to nematodes is expressed in a similar range to what’s seen in the donor tomato, that’s going to be considered as a change carrying a very low risk of unintended consequences.

On the other hand, when the genes to produce Cry proteins for insect resistance were first introduced to potatoes (never commercialized) and then corn, those were coming from the soil bacteria Bacillus thuringiensis, or Bt, a commonly used organic insecticide. While there was no reason to believe Cry proteins wouldn’t be digested by humans as any other protein, there was no history of safe use to draw on, so beyond compositional analysis, extensive animal feeding trials were required to establish safety. Any potential environmental impacts were assessed separately by the EPA under the rules for pesticide use.

(For more on Bt safety read: “I don’t want to eat food that makes insects stomachs explode! / I don’t want to eat food that’s been bred to withstand being drenched in toxic herbicides”)

What about patents?

Let’s go back to that provocative question, “How can something be substantially equivalent and yet novel enough to warrant a patent?”.

I hope at this point, the question has mostly been answered. A novel, useful characteristic or set of characteristics brought about by the labor of a breeder (or breeding program) can fall well within the normal compositional range of … let’s stick with tomatoes, and still be distinct enough to warrant the protection of patent monopoly so that the breeder’s work can be protected and paid for.

Let’s look at one example from the Johnny’s Selected Seeds catalogue: Valentine – Organic (F1) Tomato Seed. According to the catalog:

Massive early yields, deep red color, and unusually rich flavor. Developed in collaboration with Dr. Majid Foolad of Penn State University, Valentine marries the best of wild-type tomato genetics with flavorful high-performing strains. This vivid red, high-yielding, and early blight resistant tomato is the first commercial variety developed with Penn State’s patented high-lycopene breeding lines. Excellent color contrast when mixed with Nova or Golden Sweet. … U.S. Patent #8,524,992. Intermediate resistance to alternaria (early) blight.

We can look up U.S. Patent #8,524,992 and see what the breeders claim in unique and useful in their invention.


The present invention discloses new and distinct inbred tomato lines and hybrids of tomato (Lycopersicon esculentum) with higher than average lycopene content. The invention discloses publically available genetic markers which are linked to various QTLs which are associated with increased lycopene content in plants.

The present invention also discloses methods of making and using such inbred lines and hybrids. In one embodiment, the present invention discloses a new and distinct inbred tomato line, designated PSU high lycopene cherry tomato, PSCH-2; PSU high lycopene grape tomato, PSGR-23; and PSU high lycopene plum tomato, PSPL-1.

Even heirlooms can be patented if the breeder adds value with novel improvements. The breeder Jimmy Williams holds one on the Goose Creek heirloom tomato

The invention relates to a distinct tomato line with dual ripening stages, which is designated the Goose Creek tomato line. Additionally, the invention relates in part to seed and fruit of the Goose Creek tomato line as well as methods of producing the plant, fruit, and seed of the Goose Creek tomato line and Goose Creek-derived tomato plants, fruit, and seed.

So, there really is no tension between patenting plants which are also found by the FDA to be substantially equivalent.

Generally, when were are talking about objections to biotech crops and intellectual property protections, we aren’t really talking about patents, but the technology agreements that farmers sign in order to use the seeds they purchase, the way you agree to a EULA when you purchase software. But that’s a story for another day.

Read more:
GMOs: An Introduction
• About Those Industry Funded GMO Studies . . .
Answering the 3 Most Common Internet Objections to GMOs

[Please consider supporting GMO Building Blocks with an  ongoing contribution of $1, $2, $3, $5 or $10 a month on Patreon. Or make a one time donation via PayPal. ]

Please consider supporting GMO Building Blocks with an ongoing contribution of $1, $2, $3, $5 or $10 a month on Patreon.