Invasive cane toads might one day be controlled with CRISPR gene drives. Photo Credit: U.S. Geological Survey

A note from the authors: With this guest blog post we want to share the key features of an innovative method for the high-precision genome editing of wild populations that has been outlined by our team at the Wyss Institute, Harvard Medical School, and the Harvard School of Public Health. Our technical description of the proposed method was published today in eLife, while an accompanying essay on regulation and governance was published today in Science. We aim to introduce the technologywell in advance of any concrete implementationin order to start a public conversation on how we might collectively explore ways to responsibly develop and use it for the betterment of humanity and the environment.

Genome engineering technologies have revolutionized genetics, biotechnology, and medical research. We may soon be able to alter not just domesticated species, but entire wild populations and ecosystems. Why, when and how might we use these novel methods to reshape our environment?

The story begins with a new technology that has made the precise editing of genes in many different organisms much easier than ever before. The so-called “CRISPR” system naturally protects bacteria from viruses by storing fragments of viral DNA sequence and cutting any sequences that exactly match the fragment. By changing the fragments and delivering the altered system into other organisms, we can cut any given gene. If we also supply a DNA sequence that the cell can use to repair the damage, it will incorporate this new DNA, precisely editing the genome. When performed in the cells that give rise to eggs or sperm, these changes will be inherited by future generations. Because most altered traits don't improve and may even decrease the organism's ability to survive and reproduce, they generally can't spread through wild populations.

Gene drives

Genes can sometimes gain a fitness advantage that doesn't rely on the organism. Most genes in sexually reproducing species have a 50% chance of being inherited by each offspring. But many genes in nature have evolved ways to beat the odds: by being passed on more than half of the time, they gain an evolutionary advantage. This can allow an inheritance-biasing “gene drive” that is initially present in a single individual to spread over many generations until it is present in all members of a population.

One type of gene drive influences inheritance by copying itself onto chromosomes that previously lacked it. When an organism inherits such a gene drive from only one parent, it makes a cut in the chromosome from the other parent, forcing the cell to copy the inheritance-biasing gene drive—and any adjacent genes—when it repairs the damage (see the figure and additional details below).

Now imagine we want to edit a particular gene—say, one controlling the immune response of mosquitoes to malaria. We could make a mosquito with an edited version of this gene and insert the CRISPR system right next to it along with a fragment directing CRISPR to cut the original—but not the edited—gene. When our altered mosquito mates with a wild mosquito, the offspring will inherit one edited and one normal copy. CRISPR will then cut the normal copy and the cell will attempt to repair the cut by copying the edited version and the CRISPR system. The offspring will now have two copies of the edited version plus CRISPR.

This insect will mate with other insects in which the same process of turning the normal genes into edited genes will be repeated. Given enough generations, CRISPR will spread the edited gene through the entire population of mosquitoes—and this is key—even if the edited gene reduces the odds that each mosquito will reproduce.

Since CRISPR can be directed to cut essentially any gene at a precisely determined location and works in every organism we've tested, CRISPR gene drives may allow us to spread nearly any type of genome alteration through many sexually reproducing populations. We describe this possibility in detail in a paper published today in the journal eLife.

Because CRISPR itself is so precise, we can envision a number of safeguards. Alterations can be reversed by releasing a new drive with an updated version of the change. It's effectively a slow-motion “undo” button for genome alterations and could work on any type of change. Similarly, only populations that have the sequence targeted by CRISPR can be altered by a drive, potentially allowing us to target subpopulations with unique sequences. This also means that deliberately altering the sequences needed by another drive can provide protection against it, allowing us to immunize populations against specific drives and their associated changes.


There are fundamental limits to what gene drives can do. First, they cannot under any circumstances affect species that reproduce exclusively asexually, because there is no inheritance to bias. Viruses and bacteria are not susceptible, while many plants, fungi, and some animals will be resistant because they frequently reproduce without sex.

Second, gene drives can make only temporary changes. We can use them to spread traits that we consider desirable even if they are deleterious to each organism, but natural selection will eventually undo our best efforts if we give it long enough.

Third and most importantly, gene drives require many generations to spread. We could alter entire populations of fast-reproducing insects in a couple of years—depending on how many we release—but it would take decades or centuries for long-lived organisms. That's why gene drives won't be able to affect human populations without taking centuries. They're also easily detected by genome sequencing and can't spread accidentally through populations in which mate choice is artificially controlled, which greatly limits their potential to affect crops and domesticated species.


The ability to manage ecosystems by altering wild populations will have profound implications for our relationship to nature. Selective breeding and genome engineering have in many ways defined agriculture, human living and medicine, but have had comparatively little impact on most ecosystems due to the inability of domesticated crops and animals to survive in the wild. With CRISPR gene drives, we may be able to directly alter the traits or influence the population size of many non-domesticated species, which constitute the vast majority of key players in ecosystems worldwide. Given the importance of ecosystem integrity and vitality to human flourishing and the balance of life on our planet, the availability of these techniques will come with tremendous responsibility. The decision of when and where to apply this technology, and for what purposes, will be in our collective hands.

Potential Applications

Why and how might we use gene drives to intervene in a particular ecosystem? Our earlier example is perhaps the most compelling: we might use gene drives to control malaria by altering Anopheles mosquitoes that transmit the disease. Anti-malarial medicines and insecticides are losing effectiveness due to evolving resistance, while a vaccine remains out of reach despite intense research and investment. Gene drives, in contrast, might spread genes conferring malaria resistance through the mosquito populations with few if any effects on other species. Alternatively, they might be able to reduce or even eliminate the mosquitoes for long enough to permanently eradicate the malaria parasite. Similar strategies could work for other organisms that spread disease.

Gene drives might directly benefit biodiversity by controlling populations of environmentally damaging invasive species such as rats, cane toads, or lionfish. Unlike most of our current chemical or biocontrol methods, they would be specific to the target species and might be able to fully eradicate invasive populations. They might also promote more sustainable agriculture by controlling insect pests and reversing herbicide resistance in weeds, thereby supporting no-till farming.


Ecological changes caused by gene drives will be overwhelmingly due to the particular alteration and species, not by the CRISPR drive components. That means it doesn't really make sense to ask whether we should use gene drives. Rather, we'll need to ask whether it's a good idea to consider driving this particular change through this particular population. While gene drives could tremendously benefit humans and the environment if used responsibly, the potentially accessible nature of the technology raises concerns about the risks of accidental effects or even intentional mismanagement. In a new paper published in Science, we specifically address the regulation and risk governance of gene drive applications to promote responsible use.

The recent ecological concept in ethics may provide a framework for analysis and decision-making that can accommodate for a broad spectrum of moral values and worldviews.


The ability to alter populations and ecosystems using gene drives is around the corner—but not yet here today. As scientists and bioethicists, we have a professional obligation to inform society of the potential consequences of our work as early as possible. We judged the eventual development of RNA-guided gene drives to be inevitable due to the landmark theoretical work by Austin Burt, who first described the possible uses of endonuclease gene drives more than a decade ago, and the rapid advancement of CRISPR-based genome editing. After extensive discussions with experts in many fields, we elected to publish our findings in order to provide time for informed public discussion, regulatory review, and the establishment of guidelines for the safe development of the technology.

Because we are all affected by the state of our ecosystems, public oversight of technologies capable of ecological management will be essential. We recommend that all future research involving gene drives and other technologies capable of altering populations and ecosystems be conducted in full public view, with all empirical data and predictive models freely and openly shared with the global community in a transparent and understandable format. Only through broadly inclusive and well-informed public discussions can we as a society decide how best to manage our shared environment. We hope that many of you will join.

George Church is a member of Scientific American's Board of Advisors


How Endonuclease Gene Drives Work

Gene drives for ecosystem management (click to enlarge)


Standard drives spread introduced genes or specific patterns of genome changes through populations. They might be used to drive a gene that interferes with the spread of a disease, such as an anti-malarial peptide, or to disrupt a natural gene important for disease spread.

Suppression drives reduce the number of organisms in a target population. Naturally occurring suppression drives are always found together with a resistance element that allows the species to survive. It is believed that the emergence of drives in the past may have driven species all the way to extinction when resistance did not develop in time. There are many different ways to build a suppression drive, some of which have quite different effects. They could be used to control the populations of environmentally destructive invasive species such as mosquitoes, rats, cane toads, or lionfish that currently threaten many ecosystems. They might even be able to render pest populations uniquely vulnerable to molecules that don't affect other organisms.

Reversal drives undo earlier genome changes caused by natural evolution, human-inserted transgenes, earlier gene drives, or even naturally evolved mutations such as those conferring pesticide or herbicide resistance in pests and weeds.

Immunization drives immunize existing populations to prevent them from being affected by other gene drives.

Precision drives can only spread alterations through populations of organisms that have a unique DNA sequence.