Scientists can't agree on how to test our future foods

In 1994, our grub changed forever — the FDA approved a genetically engineered food for the first time, and soon after a lab-designed tomato called the Flavr Savr made its way to grocery store shelves.

It included a gene originally found in bacteria, along with a modified tomato gene. Their tweaks messed with the production of an enzyme that causes the fruit to rot, so it remained relatively firm for longer than conventional varieties

In the following years, giant agricultural companies like Monsanto began rolling out new types of GM foods (which are now officially referred to as "bioengineered”). After all, it’s profitable to sell seeds with customized genomes that can resist herbicides, fight off insects, and even enhance vitamin content, among other helpful qualities. Today, more than 90 percent of U.S. corn, upland cotton, and soybeans are genetically engineered.

Tinkered with plants is nothing new. We have evidence of people “engineering” crops since around 7800 B.C.E. by selecting for varieties with ideal traits.

In recent decades, though, we have gotten more precise with our custom plants. The Flavr Savr and most of its successors were created from recombinant DNA technology, a method developed in the 1970s that inserts genes from another plant, or even a different species, into a crop to create desirable characteristics like higher yields and drought tolerance.

Today, the method is being eclipsed by CRISPR, a genome-editing technology that’s more precise: It can snip at precise sections of DNA to achieve desired traits in crops. But unlike past tried-and-true methods, we’re less sure of CRISPR’s unintended consequences in crops.

While scientists have pointed out that gene edits play out identically to natural mutations in food, and they may even transform less frequently than in the wild, the powerful technology’s “off-target” impacts aren’t yet entirely understood due to limited data.

Regardless of the technique employed, companies (somewhat) voluntarily test to make sure that new GMOs won’t cause allergic reactions or other health consequences like organ failure, and they also look out for possible environmental hazards, such as the creation of stubborn new weeds.

“None of these regulations are really fit for purpose … They’re not scientifically defensible.”

But according to a new paper published last week in the journal Science, a group of agricultural scientists, biologists, and other experts say current testing procedures aren’t enough to ensure that both people and our surroundings are safe when new GM crops hit our shelves.

“None of these regulations are really fit for purpose … They’re not scientifically defensible,” says Fred Gould, an evolutionary biologist at North Carolina State University and co-author of the new paper.

Stricter testing procedures could also help earn the public’s trust, the authors claim, which is sorely lacking. More than a third of U.S. respondents said GMOs are generally unsafe to eat in a Pew Research Center survey taken between October 2019 and March 2020. Meanwhile, about 90 percent of scientists think they’re safe to consume.

In fact, researchers are betting on genetically modified crops to feed the world’s growing population and keep producing meals amid the worsening impacts of climate change. Something has to give.

The paper suggests a solution: Companies should start off by scanning plants’ genomes — similar to what we’ve done for yeast, humans, frogs, dogs, and a bunch of other creatures — to find tiny molecular changes and glean useful genetic information. But not all scientists agree that the time-intensive method would be worth it due to the added costs and potential production delays.

Sniffing out risks

While there’s no law that dictates how companies test genetically modified foods, the FDA runs a consultation program that often requires heaps of testing before a GMO food can enter the market. “It is technically voluntary, but the fact is that the marketplace will not touch the product if it has not been past the FDA,” Wayne Parrott, a plant geneticist at the University of Georgia, tells Inverse.

And as for plants created to resist pests, the law gives the Environmental Protection Agency the right to thoroughly study them.

Lab tests can confirm that the intended changes in a plant are safe and also avoid unintended shifts that can produce harmful substances in food. The latter requires the bulk of the testing, Parrott says.

To check for hazards, companies take a close look at the proteins produced by the modified genes and see whether they could be toxic or cause allergic reactions in people. They also check if people are easily able to digest these proteins and make sure that the nutrient levels are comparable to those in the original crop.

It’s becoming more common to sequence the modified genes to indicate toxicity, he says. But if that’s not enough to get a clear answer, they can feed high doses of the isolated genes to mice to test for short-term dangers. To assess long term-effects, labs give a series of increasing doses to rats and examine their organs for signs of damage.

Overall, tests very rarely find anything wrong with GM crops, Parrott says, and if they do, they’re never sold in stores. “These days, any gene that might have even a small chance of being allergenic or toxic is discarded up front and never put into a pipeline,” he explains.

Playing by new rules

But over the past few years, certain safety standards have shifted. In 2020, for example, the government relaxed the rules by exempting some gene-edited plants from regulation. For instance, it now offers automatic approval for small shifts, such as adjustments for certain climates, on GM crops that have already been proven safe.

And if researchers concoct a new plant that could have theoretically been bred through conventional methods, no testing is required. This rule assumes that changing one base pair of DNA, or deleting any size of genetic material, could occur naturally. But even a single base pair change could, for example, trigger the production of a harmful chemical compound, the recent Science article claims.

Parrott doesn’t think this rule shift poses a risk. After all, toxins spawned by unintended changes “have never been found, at least not in the mainstream literature that uses sound methodology,” he says.

It does worry Gould, the co-author of the new paper. He and his colleagues say that the government is blurring the line between “conventional” and “non-conventional” methods.

After all, regulators don’t even explicitly define what conventional means, and it can include newer technologies such as genomic selection, in which giant databases chock-full of genetic information are used to guide the development of new crops.

“It’s cheaper than doing safety testing, and it’s scientifically supportable.”

Instead of basing testing on the precise method used to forge a new type of genetically modified crop, the paper proposes a universal approach that scans new crop varieties for unexpected DNA changes that could cause trouble. This will be particularly helpful as new technologies we can’t even imagine today emerge in the future, Gould explains.

The research group recommends a variety of “-omics” methods, such as transcriptomics, which examine the RNA molecules that produce important proteins, and metabolomics, which zooms in on the compounds produced by plants. All in all, they say these techniques could snap a detailed picture of brand-new plant varieties.

If the new product is flagged for a novel molecular characteristic that could risk human health or the environment, or if the differences can’t clearly be interpreted in the lab, the company can then perform safety testing. Overall, this framework wouldn’t trigger additional analysis for most new varieties, the paper notes.

This process could offer important insight without getting bogged down by the specifics of genetic modification methods — which may be particularly helpful as radically new technologies continue to emerge, according to Gould.

“It’s cheaper than doing safety testing, and it’s scientifically supportable,” he says.

A waste of time?

Gould’s paper highlights the significant challenges in regulating GM crops, says Robert Stupar, a plant geneticist at the University of Minnesota.

“I think the authors did a nice job of highlighting some of the issues with how products with plant biotechnology are regulated these days,” Stupar says. “I really appreciate their brainstorming on this subject because it’s complicated.”

However, Stupar adds, novel segments that pop up in manipulated plant DNA don’t necessarily indicate potential risk. For example, the grain has plenty of natural variations to begin with: Two conventionally bred lines of corn could have millions of genetic differences, and new mutations occur naturally in every generation.

“I think there’s the potential that this could lead to overregulation, and they do mention that in the article,” he notes.

Scientists will likely argue that this suggestion could stifle innovation.

The paper also doesn’t define the specific threshold of observed changes that will prompt safety testing, but Gould says that definition is up to further research and public input. “We’re proposing this as a method and it requires insights from society to determine what those levels are,” he says.

He compares it to the evolving standards for safe levels of forever chemicals in drinking water recognized by the government, which officials have determined based on mounting research (and public pressure).

But as with other strict regulations, scientists will likely argue that this suggestion could stifle innovation. “Requiring -omics as a preliminary screen in the absence of a risk hypothesis needlessly adds to development costs and will be a barrier to public labs and [less profitable] crops, without any increase in safety,” Parrott says.

Still, Gould thinks that his proposed approach could help combat misinformation around GMOs by giving consumers more of a say in what ends up on their plates. “You have to do it in a way that’s not just the companies rolling it out and nobody trusting them,” he says.