This article was originally published on GM Watch on August 9th, 2023

Uncompromising review is published as European Commission and UK government continue pretence that gene editing is precise, predictable, and safe. Report: Claire Robinson

The latest in a long series of papers has been published, detailing unintended effects of CRISPR gene editing. The new review summarises the many types of serious unintended on-target (at the intended edit site) DNA damage resulting from CRISPR/Cas gene editing.

The review appears as the European Commission and the UK government maintain their pretence that gene editing is a precise, predictable, and controllable technique and that food plants made with this technology are therefore as safe as those produced by conventional breeding.

The authors of the new paper, from Rice University in the US, reviewed the literature on CRISPR gene editing in human, primate, and mouse cells. They found that CRISPR-induced double-strand breaks in the DNA caused numerous large unintended on-target genetic damages, including large and small deletions and insertions, and chromosomal rearrangements of genetic material. And they note that even large on-target gene modifications are not detectable by standard methods.

Because the unintended effects of CRISPR gene editing highlighted in the new review are on-target mutations at the intended edit site, improving the targeting of the editing tool isn’t going to solve these problems, as GMWatch has warned before.

CRISPR’s “potentially irreversible safety risks”

As we’ve come to expect, the authors of the new review are from the clinical gene therapy arena rather than the agricultural genetic engineering arena. As an example of the widespread concern about CRISPR’s imprecision in the clinical arena, a recent article in Forbes warns that gene editing “comes with potentially irreversible safety risks”.

In stark contrast, plant and livestock animal gene editing scientists seemingly prefer to pretend that there are no unintended effects of CRISPR gene editing. This lie of omission leads to the grotesque and ironic situation where conflicted scientists and misguided politicians parrot sales pitches about precision and predictability, even as the scientific evidence of the opposite mounts up.

Inadequate screening for unintended effects

The authors of the new paper draw attention to the inadequate methods that many GMO developers use to look for unintended effects of gene editing. While they state that on-target large gene modifications can occur at high frequencies, they point out that they are “undetectable by standard short-range PCR based assays, leading to data misinterpretation, reduced efficacy, and potential safety concerns in therapeutic gene editing”.

GMWatch has repeatedly drawn attention to the biased and inadequate screening methods generally used by genetic engineers in the agricultural arena, which means they will miss many unintended effects.

As a result of this “don’t look, don’t find” approach, the sector has successfully pulled the wool over the eyes of the European Commission, which is attempting to remove regulatory safeguards from most gene-edited GMOs. GMO developers have done a similar whitewash job for the UK government, which has passed legislation, laughably called the Genetic Technology (Precision Breeding) Bill, which wrongly assumes that gene-edited plants are no more risky than conventionally bred ones.

The risks and problems of agricultural gene-edited GMOs have been emphasised by scientists in many peer-reviewed articles, so at this point there is no excuse for remaining ignorant.

The lie of cleaning up gene-edited genomes

Where GMO proponents admit to the unintended effects of gene editing, they insist that they can easily be cleaned up by subsequent cross-breeding.

But it’s not possible to clean up a genome from unintended effects when you don’t know what they are. And it’s clear that agricultural GMO developers are not looking properly for such effects. They use the wrong analytical tools, so it’s equivalent to looking for a lost set of keys in the dark without any form of night vision.

In addition, an analysis by Dr Yves Bertheau, director of research at the French National Institute for Agricultural Research, shows that “cleanup” cross-breeding, where it is carried out, is limited by developers’ habits, expertise, and the available material. So in most cases, “genome cleaning” by backcrossing is not done thoroughly enough. As a result, unintended genetic damage can remain in the final marketed GMO, with potentially dangerous consequences.

We conclude that GMO proponents, in claiming that they will “clean up” their genetic error-prone GM plants, are grossly misleading politicians and the public.

Prime editing and base editing won’t save the technology

Adaptations of CRISPR gene editing methods known as base editing and prime editing are often touted as safer alternatives to standard CRISPR/Cas gene editing, as the latter causes double strand breaks in the DNA which often result in imprecise and unpredictable repairs by the cell’s own repair mechanism. Base editing and prime editing aim to cause single strand breaks in the DNA instead of double strand breaks and are therefore less prone to DNA repair leading to large deletions, insertions, and rearrangements of genetic material.

But the authors of the new paper say that base editing and prime editing have problems too, as they can also introduce a “low level” of double strand breaks and may introduce unintended large deletions of DNA. They say scientists need “a better understanding” of the unintended effects of these techniques before they can be used in gene therapy.

Expert comment on the new paper

Commenting on the new review, London-based molecular geneticist Prof Michael Antoniou called it “very telling”, as it “shows that all the hype at the beginning about how cheap, simple and easy CRISPR gene editing is no longer applies. You are having to add so many modifications to the basic machinery that simplicity has gone out of the window. The best that’s achievable is a reduced frequency of off-target and unintended on-target mutations that result in unintentional disruption to one or more gene functions. It will be difficult or impossible to eliminate them because we are not in full control of the biological processes involved.”

Prof Antoniou continued, “The authors state that base editing and prime editing can result in double strand breaks in the DNA, albeit at a low frequency. But in gene therapy, it only takes one double strand break and inadvertent disruption of a crucial gene function, such as a cancer suppressor gene, to send a cell down the cancer pathway.

“It has become clear from the accumulating evidence that in terms of unintended outcomes, gene editing appears no better than gene addition gene therapy. In gene addition gene therapy, you can have very low frequency insertional mutagenesis [DNA damage caused by insertion of the introduced therapeutic gene], with a possible cancer outcome. In this context, gene editing could even be more risky than gene addition.

“In general, the clinical field has strict safety checks. But we hope that with gene therapy using CRISPR gene editing, developers are conducting carcinogenicity studies in cell cultures, as well as long-term studies of at least a year’s duration in animals, such as mice. Only such long-term studies could show cancers that in humans might take some years to appear. In short, gene editing needs the same rigorous testing as is applied to gene addition therapy, to determine the risk-benefit balance.”

Relevance of the paper to agricultural gene-edited GMOs

Prof Antoniou explained how the new paper is relevant to agricultural CRISPR gene-edited GM plants: “The mutations that the review flags up – large deletions and rearrangements at the intended on-target site – can only comprehensively be detected if the developer does ultra-deep, preferably long-read, whole genome sequencing. In the absence of this analysis, they will miss many of these genetic errors.

“We know they are not doing this, for the most part. In general, all they do is short-range PCR analysis at the intended edit site and at locations of the genome that computer programs predict could be affected by off-target CRISPR activity. However, it is known that short-range PCR analysis will miss unintended large-scale deletions, insertions and rearrangements of DNA. Minimally they need to undertake long-range PCR to try to capture the occurrence of large unintended genetic alterations.

“This is especially needed for plants with more than two sets of chromosomes (polyploid). For example, bread wheat has six sets (hexaploid). When you gene edit this wheat, you are trying to target all six sets of chromosomes. You may have one gene locus edited in the way you want, but what about the five others? Regulators should demand proof that this is being addressed properly by plant gene editors.

“Such analyses are crucial because unintended large genetic rearrangements resulting from CRISPR gene editing could affect the function of many genes, potentially resulting in a toxic or allergenic plant, or one that has adverse effects on the environment.

“Regarding CRISPR gene-edited animals, the kinds of genetic errors noted in the new review could cause health problems such as cancer or birth defects – as would apply to humans.”

As far as base and prime editing in agricultural plants is concerned, the range of traits that can be obtained using these techniques is likely to be very limited, compared to using CRISPR/Cas to create double strand breaks. And while unintended on-target and off-target (at locations in the genome other than the target edit site) mutations may be reduced, based on the evidence presented in the new review, they will not be eliminated. And again, no one appears to be looking properly.

In GMWatch’s view, even though many politicians and regulators have been fooled by claims of precision and predictability, the unavoidable frequent and genome-wide unintended genetic damage caused by the CRISPR gene editing process as a whole (a combination of plant tissue culture, plant cell transformation and action of the CRISPR gene editing tool) means that it will not only fail to deliver on its promises, but runs a high risk of causing lasting harm to health and the environment.

The new review:

Park, S. H., Cao, M., and Bao, G. (2023). Detection and quantification of unintended large on-target gene modifications due to CRISPR/Cas9 editing. Current Opinion in Biomedical Engineering, 100478.


* CRISPR/Cas9 systems generate unintended large on-target gene modifications.
* Large deletions, insertions, and chromosomal rearrangements can have high frequencies.
* Large gene modifications are not detectable by standard short-range PCR based assays.
* Novel methods have been developed for analyzing large on-target gene editing outcomes.
* Comprehensive understanding of on-target editing outcomes is critical to improve safety.


CRISPR/Cas9 based gene editing typically functions by creating a DNA double-strand break (DSB) at the intended target locus in a cell. Recent reports showed the occurrence of unintended on-target large gene modifications by CRISPR/Cas9-induced DSB, including large deletions, insertions, and chromosomal rearrangements, in addition to small insertions and deletions. These on-target large gene modifications can have high frequencies, undetectable by standard short-range PCR based assays, leading to data misinterpretation, reduced efficacy, and potential safety concerns in therapeutic gene editing. Here, we summarize the recent advances in analyzing large on-target gene editing outcomes and their implications to clinical application, and discuss opportunities for future improvements.