Truth: Genetic engineering and mutation breeding are both risky and should be strictly regulated

Myth at a glance

GM proponents often compare GM with radiation- or chemical-induced mutation breeding (mutagenesis) and claim that these methods are even more mutagenic than GM and at least as disruptive to gene expression. They argue that crops developed by mutation breeding are widely viewed as safe and have not caused health problems; and that therefore GM crops should not be subjected to stricter regulation than mutation-bred crops.

Some GM proponents imply that mutagenesis is equivalent to conventional breeding.

However, while mutagenesis is used in conventional breeding, mutation breeding is not the same as conventional breeding. Mutation breeding is unpredictable and risky, and crops produced in this way should be as strictly regulated as GM crops.

GM proponents often compare GM with mutation breeding (or mutagenesis), which they say has been used for decades in conventional plant breeding and is not controversial. They argue that mutation breeding is used by conventional plant breeders and that mutation-bred plants have a history of safe use and do not cause ill health.1 GM proponents also say that genetic modification is more precise than mutation breeding, and imply that therefore, GM plants should not be regulated any more strictly than those produced by mutation breeding.2

However, these arguments are flawed, for the reasons explained below.

What is mutation breeding?

The physical form of an organism’s genetic blueprint is the sequence of the four “bases”, or “letters” (A, G, C, T) of the genetic alphabet. The sequence in which these four “letters” are linked together to form the DNA molecule determines the information contained in that molecule, just as the sequence in which the 26 letters of our alphabet are placed on this page determines its information content.

You can change the meaning of a sentence by changing the sequence of letters in the sentence, and you can change the “meaning” of a gene or its associated genetic control elements by changing the sequence of letters within the genetic code of that gene or control element. Mutations are physical alterations in the sequence of the four letters of the genetic alphabet within the DNA.

Mutation breeding is the process of exposing plant seeds to mutagens – physical or chemical agents that damage the DNA, causing mutations. In practice, these agents are either ionizing radiation (X-rays or gamma rays) or compounds that physically or chemically react with DNA.

The types of mutations that can be created range from a change in a single genetic letter (for example, “A” can be replaced with “C”, or “G” with “T”), to the deletion of one or many letters, to rearrangements of small or large stretches of the DNA sequence.

This process of change in the DNA is known as mutagenesis. Mutagenesis can completely destroy the function of a gene – “knockout” its function – or it can cause the gene to direct the cell to produce one or more proteins with altered function. In addition, mutagenesis can alter the functioning of the genetic control elements associated with a gene or genes and thus alter the amount, timing, or location of the protein products produced from them. The resulting plant is called a mutant.

It is a fortuitous and infrequent event when a mutation improves the functioning of an organism. More often, mutations are damaging or silent (no observable effect). Damage can range from death of the plant, to minor reductions in productivity or vigour, to changes in the function or structure of the organism, and even to the quality or safety of the food derived from the crop plant.

Once plants carrying radiation-induced mutations have been created, they are crossed with other crop varieties using conventional breeding (the same process is used with GM crop varieties). However, mutation breeding is not in itself conventional breeding.

Where did radiation-induced mutation breeding come from?

Mutation breeding using radiation started in the 1920s. It became more widely used in the 1950s, after the US atomic bombing of Japan at the end of World War II in 1945. In the wake of the devastation, there was a desire to find uses for the “peaceful atom” that were helpful to humanity. Atomic Gardens were set up in the US and Europe, and even in Japan, with the aim of creating high-yielding and disease-resistant crops. They were laid out in a circle with a radiation source in the middle that exposed plants and their seeds to radiation. This caused mutations in the plants, which radiation enthusiasts hoped would be beneficial. Public relations campaigns euphemistically described the plants as “atom energized”.

However, the results of these projects were poorly documented and do not qualify as scientific research. It is unclear whether any useful plant varieties emerged from Atomic Garden projects.3

Today, radiation-induced mutation breeding is carried out in laboratories. This branch of plant breeding retains strong links with the nuclear industry. The only database of crop varieties generated using radiation- and chemically-induced mutation breeding is maintained by the UN Food and Agriculture Organization in partnership with the International Atomic Energy Agency.4 Many studies and reports that promote radiation-induced mutation breeding are sponsored by organizations that also promote nuclear energy.56

Is mutation breeding widely used?

Mutation breeding is not a widely used or central part of crop breeding. It is a minor footnote to the advances that conventional breeding has brought to agriculture, although a handful of crop varieties have apparently benefited from it. The database maintained by the UN Food and Agriculture Organisation and the International Atomic Energy Agency keeps track of plant varieties that have been generated using mutation breeding and cross-breeding with a mutant plant.4 The database contains only around 3,000 such plant varieties, and this number includes not only food crop plants but also ornamental plants.7 It also includes not only the primary mutant varieties generated through mutagenesis, but also any varieties that have been created by crossing the primary mutant varieties with other varieties by conventional breeding. Thus the actual number of primary mutant varieties is a fraction of the 3,000 varieties listed in the database.

Conventional breeding, in contrast, has produced millions of crop varieties. The Svalbard seed vault in the Arctic contains over 770,000 seed varieties.8 In 2009 its seed stocks were estimated to represent one-third of our most important food crops.9 So quantitatively speaking, mutation breeding has proved to be of only marginal importance in crop development.

Why isn’t mutation breeding more widely used?

The process of mutagenesis is risky, unpredictable, and does not efficiently generate beneficial mutations. Studies on fruit flies suggest that about 70% of mutations will have damaging effects on the functioning of the organism, and the remainder will be either neutral or weakly beneficial.10

Because of the primarily harmful effects of mutagenesis, living organisms have DNA repair mechanisms to correct mutations and minimize their impacts. The primarily harmful effect of mutations is reflected in the policies of regulatory agencies around the world, which are designed to minimize or eliminate exposure to radiation and other manmade mutagens.

In plants as well as fruit flies, mutagenesis is a destructive process. One textbook on plant breeding states, “Invariably, the mutagen kills some cells outright while surviving plants display a wide range of deformities.”11 Experts conclude that most such induced mutations are harmful and lead to unhealthy and/or infertile plants.11,12

A report by the UK government’s GM Science Review Panel concluded that mutation breeding “involves the production of unpredictable and undirected genetic changes and many thousands, even millions, of undesirable plants are discarded in order to identify plants with suitable qualities for further breeding.”13

Occasionally, mutagenesis may give rise to a previously unknown feature that may be beneficial and can be exploited. Commercially useful traits that have emerged from mutation breeding include the semi-dwarf trait in rice, the high oleic acid trait in sunflower, the semi-dwarf trait in barley, and the low-linolenic acid trait in canola (oilseed rape).7,14,15 It is interesting to note that all of these traits are the result of destruction of the function of one or more natural genes, not the remodelling or fine-tuning of genes or the proteins they encode. This reflects the brute-force nature of the mutation breeding technique.

The process of screening out undesirable mutants and identifying desirable ones for further breeding has been likened to “finding a needle in a haystack”.11 The problem is that only certain types of mutations, such as those affecting shape or colour, are obvious to the eye. These plants can easily be discarded or kept for further breeding as desired. But other more subtle changes may not be obvious, yet nonetheless can have important impacts on the health or performance of the plant. Such changes can only be identified by expensive and painstaking testing.11

In retrospect, it is fortunate that mutation breeding has not been widely used because that has reduced the likelihood that this risky technology could have generated crop varieties that are toxic, allergenic, reduced in nutritional value, vulnerable to pests or environmental stressors, or harmful to the environment.

Why worry about mutations caused in genetic engineering?

GMO proponents make four basic arguments to counter concerns about the mutagenic aspects of genetic engineering.

1. “Mutations happen all the time in nature”

GMO proponents say that mutations happen all the time in nature as a result of various natural exposures, for example, to ultraviolet light, so mutations caused by genetic engineering of plants are not a problem.

In fact, mutations in nature are a low-frequency event.7 And comparing natural mutations with those that occur during genetic modification is like comparing apples with oranges. Every plant species has encountered environmental mutagens, including certain types and levels of ionizing radiation and chemicals, throughout its natural history and has evolved mechanisms for preventing, repairing, and minimizing the impacts of any mutations caused. But plants have not evolved mechanisms to repair or compensate for the insertional mutations that occur during genetic modification. Also, the high frequency of mutations caused by tissue culture during the process of developing a GM plant is likely to overwhelm the plant’s repair mechanisms.

Homologous recombination events that move large stretches of DNA around a plant’s genome do occur in nature. But the mechanisms of homologous recombination are very precise, and rarely cause mutations. Also, the DNA sequences that undergo rearrangement during homologous recombination are already part of the plant’s own genome, not DNA that is foreign to the species.

In addition, if mutations were to occur that compromised the quality of the food produced by the plant, for instance, by producing unexpected toxins, the long co-evolution process between humans and their food crops would have enabled such harmful mutants to be eliminated from the breeding process.

2. “Conventional breeding is less precise and more disruptive to gene expression than GM”

Some GMO proponents cite a study by Batista and colleagues16 to argue that chemical- or radiation-induced mutagenesis, used in “conventional” breeding, is less precise and more disruptive to gene expression than GM. They term radiation-induced mutagenesis “conventional radiation treatment” and argue on the basis of papers discussing mutation-bred crops that “conventional plant breeding causes mutations” – appearing to imply that mutation breeding is synonymous with conventional breeding. They add that plants developed in this way are widely accepted and have not caused ill health in consumers.1

However, such arguments misrepresent the study of Batista and colleagues and the nature of conventional breeding and mutation breeding. Batista and colleagues did not compare conventional breeding with GM, but radiation-induced mutation breeding with GM.16

Mutation breeding is not the same as conventional breeding. While radiation- and chemical-induced mutation breeding has been used in tandem with conventional breeding, it is not in itself conventional breeding. Mutation breeding only escaped regulation because of the widespread ignorance about the potential effects of mutations in food crops at the time that the method began to be used in crop breeding.

Batista and colleagues’ research actually provides strong evidence to support the argument that GM is highly disruptive to gene expression. The study found that in rice varieties developed through radiation-induced mutation breeding, gene expression was disrupted even more than in varieties generated through genetic modification. They concluded that for the rice varieties examined, mutation breeding was more disruptive to gene expression than genetic engineering.16

Batista and colleagues did not compare GM with conventional breeding, but compared two highly disruptive methods – genetic engineering and mutation breeding – and concluded that genetic engineering was, in the cases considered in their study, the less disruptive of the two methods.

One GM proponent nonetheless concludes, based on the Batista paper, that “the potential for harm in both cases is trivial”.2But this was not the conclusion that Batista and colleagues drew from their findings. They concluded that all crop varieties produced by either mutation breeding or genetic engineering should be subjected to safety assessment.16

We agree with the conclusions of Batista and colleagues. While their study does not examine enough GM crop varieties and mutation-bred crop varieties to enable generalized conclusions about the relative risks of mutation breeding and genetic engineering, it does provide evidence that both methods significantly disrupt gene regulation. It also suggests that crops generated through these two methods should be assessed for safety with similar levels of rigour. The fact that the risks of mutation breeding have been overlooked by regulators does not justify overlooking the risks of GM crops as well.

Significantly, an expert committee of the US National Research Council concluded that genetic engineering was more likely to cause unintended changes than all other crop development methods except mutation breeding.17

Regulations around the world should be revised to treat mutation-bred crops with the same sceptical scrutiny with which GM crops should be treated.

3.More mutations occur as a result of natural breeding than of genetic engineering”

GM proponents say that in conventional breeding, traits from one variety of a crop are introduced into another variety by means of a genetic cross. They point out that the result is offspring that receive one set of chromosomes from one parent and another set from the other. They further point out that for some genes, the maternal and paternal versions will be identical, but for many other genes, the maternal and paternal versions will be different. Thus there is the potential that the genetic makeup of the offspring will deviate from that of either parent by as much as 50%. That is, tens of thousands of the genes carried by the offspring could be different from the genes carried by one of the parents.

They suggest that the result is a patchwork that contains tens of thousands of deviations from the DNA sequence and genetic information present in the chromosomes of either parent. They imply that these deviations can be regarded as tens of thousands of mutations, and conclude that because we don’t require crop varieties resulting from such genetic crosses to undergo biosafety testing before they are commercialised, we should not require GMOs, which they claim contain only a few mutations, to be tested.

But this is a spurious argument. The versions of a gene – called alleles – contributed by both the mother and father are typically not different due to recent mutagenic events. These alleles are established versions of the gene that have survived the process of natural selection over the ages because they confer distinct, useful characteristics onto the individual that carries them.

Thus the genome and phenotype of the offspring resulting from a genetic cross of two varieties is not the result of random mutations, but of the precise combination of genetic material contributed by both parents. This is a natural mechanism operating on the level of the DNA to generate diversity within a species, yet at the same time preserve the integrity of the genome with letter-by-letter exactness.

Genetic engineering, on the other hand, is an artificial laboratory procedure that forces foreign DNA at random into the DNA of the cells of a plant. Once the engineered gene is introduced into the nucleus of the cells, it breaks randomly into the DNA of the plant and inserts into that site. This process results in at least one insertional mutation. However, other steps in the genetic engineering process generate hundreds, possibly even thousands, of mutations throughout the plant’s DNA.18

For these reasons, conventional breeding is far more precise and carries fewer mutation-related risks than genetic engineering.

4. “We will select out harmful mutations”

GM proponents say that even if harmful mutations occur, that is not a problem. They say that during the process of developing a GM crop, the GM plants undergo many levels of screening and selection and the genetic engineers will catch any plants that have harmful mutations and eliminate them during this process.1

The process of gene insertion during genetic modification selects for insertion of engineered GM gene cassettes into regions of the host (recipient) plant cell genome where many genes are being actively expressed. Insertion of GM sequences into such regions has a high inherent potential to disrupt the function of active genes native to the plant’s genome.

In some cases, the disruption will be fatal – the engineered cell will die and will not grow into a GM plant. In other cases, the plant will compensate for any disturbance in the function of genes, or the insertion will occur at a location that seems to cause minimal disruption of the plant cell’s functioning. This is what is desired. But just because a plant grows vigorously and has a healthy green colour does not mean that it is safe to eat and safe for the environment. It could have a mutation that causes it to produce substances that harm consumers or to damage the ecosystem.

Genetic engineers do not carry out detailed screening that would catch all plants producing potentially harmful substances. They introduce the GM gene(s) into hundreds or thousands of plant cells and grow them out into individual GM plants. If the gene insertion process has damaged the function of one or more plant cell genes that are essential for survival, the cell will not survive this process. So plants carrying such “lethal” mutations will be eliminated. But the genetic engineer is often left with several thousand individual GM plants, each of them different, because:

  • The engineered genes have been inserted in different locations within the DNA of each plant
  • Other mutations or disturbances in host gene function have occurred at other locations in the plants through the mechanisms described above.

How do genetic engineers sort through the GM plants to identify the one or two they are going to commercialise? They do a test that allows them to find the few plants, among many thousands, that express the desired trait at the desired level. Of those, they pick some that look healthy, strong, and capable of being bred on and propagated.

That is all they do. Such screening cannot detect plants that have undergone mutations that cause them to produce substances that are harmful to consumers or lack important nutrients.

It is unrealistic to claim that genetic engineers can detect all hazards based on obvious differences in the crop’s appearance, vigour, or yield. Some mutations will give rise to changes that the breeder will see in the greenhouse or field, but others will give rise to changes that are not visible because they occur at a subtle biochemical level or manifest only under certain circumstances. So only a small proportion of potentially harmful mutations will be eliminated by the breeder’s superficial inspection. Their scrutiny cannot ensure that the plant is safe to eat.

Some agronomic and environmental risks will be missed, as well. For instance, during the GM transformation process, a mutation may destroy a gene that makes the plant resistant to a certain pathogen or a specific environmental stress like extreme heat or drought. But that mutation will be revealed only if the plant is intentionally exposed to that pathogen or stress in a systematic way. GM crop developers are not capable of screening for resistance to every potential pathogen or environmental stress. So mutations can sit like silent time bombs within the GM plant, ready to “explode” at any time when there is an outbreak of the relevant pathogen or an exposure to the relevant environmental stress.

An example of this kind of limitation was an early – but widely planted – variety of Roundup Ready soy. It turned out that this variety was much more sensitive than non-GM soy varieties to heat stress and more prone to infection.19


Like genetic engineering, radiation-induced mutagenesis is risky and mutagenic. It is not widely used in plant breeding because of its high failure rate. Comparing genetic engineering with radiation-induced mutagenesis and concluding that it is safe is like comparing a game of Russian Roulette played with one type of gun with a game of Russian Roulette played with another type of gun. Neither is safe.

A more useful comparison would be between genetic engineering and conventional breeding that does not involve radiation- or chemical-induced mutagenesis. This is the method that has safely produced the vast majority of our crop plants over millennia and that is most widely used today. It is also far more successful. All the increases in crop yield achieved around the world in the last several decades are due to conventional breeding, not genetic engineering.


  1. Academics Review. The use of tissue culture in plant breeding is not new. 2014. Available at:
  2. Genetic Literacy Project. GMOs vs. mutagenesis vs. conventional breeding: Which wins? 2013. Available at:
  3. Anon. Atomic gardens: Interview with Paige Johnson. Pruned. Published April 20, 2011.
  4. Food and Agriculture Organization (FAO) and International Atomic Energy Agency (IAEA). Mutant variety database (MVGS). 2010. Available at:
  5. Kodym A, Afza R. Physical and chemical mutagenesis. Methods Mol Biol. 2003;236:189-204. doi:10.1385/1-59259-413-1:189.
  6. Novak FJ, Brunner H. Plant breeding: Induced mutation technology for crop improvement. IAEA Bull. 1992;4:25–33.
  7. Jain SM. Mutagenesis in crop improvement under the climate change. Romanian Biotechnol Lett. 2010;15:88–106.
  8. Ministry of Agriculture and Food (Norway). Svalbard Global Seed Vault secures future seed. 2013. Available at:
  9. BBC News. More seeds for “doomsday vault.” Published February 26, 2009.
  10. Sawyer SA, Parsch J, Zhang Z, Hartl DL. Prevalence of positive selection among nearly neutral amino acid replacements in Drosophila. Proc Natl Acad Sci USA. 17;104:6504-10. doi:10.1073/pnas.0701572104.
  11. Acquaah G. Principles of Plant Genetics and Breeding. Oxford, UK: Wiley-Blackwell; 2007. Available at:
  12. Van Harten AM. Mutation Breeding: Theory and Practical Applications. London: Cambridge University Press; 1998.
  13. GM Science Review Panel. First report: An open review of the science relevant to GM crops and food based on interests and concerns of the public. 2003.
  14. Ahloowalia BS, Maluszynski M, Nichterlein K. Global impact of mutation-derived varieties. Euphytica. 2004;135:187–204.
  15. Maluszynski M, Szarejko I. Induced mutations in the Green and Gene Revolutions. In: Tuberosa R, Phillips RL, Gale M, eds. Bologna, Italy: Avenue Media; 2003.
  16. Batista R, Saibo N, Lourenco T, Oliveira MM. Microarray analyses reveal that plant mutagenesis may induce more transcriptomic changes than transgene insertion. Proc Natl Acad Sci USA. 4;105:3640-5. doi:10.1073/pnas.0707881105.
  17. National Research Council. Safety of genetically engineered foods: Approaches to assessing unintended health effects. Washington, DC: The National Academies Press; 2004. Available at:
  18. Latham JR, Wilson AK, Steinbrecher RA. The mutational consequences of plant transformation. J Biomed Biotechnol. 2006;2006:1–7. doi:10.1155/JBB/2006/25376.
  19. Coghlan A. Monsanto’s Roundup-Ready soy beans cracking up. New Sci. 1999. Available at: