Substack cometh, and lo it is good. (Pricing)

The technology and science of “gene drive”

41eQOJU5FBL._SX331_BO1,204,203,200_For God doth know that in the day ye eat thereof, then your eyes shall be opened, and ye shall be as gods, knowing good and evil.

-Genesis 3:5

Over at MIT Tech Review my friend Antonio Regalado has an article up, We Have the Technology to Destroy All Zika Mosquitoes, which seems to promise almost god-like powers in the hands of bio-engineers. But, one thing that I did suggest was that these sorts of articles which focus on technological element of developing story don’t inform the public that there has long been an understanding that there are selfish genetic elements which violate Mendel’s law of segregation. In other words, the technology is leveraging a natural phenomenon. These forces may be pervasive, but they do not seem to ultimately upend the balance which allows for the persistence of Mendelian transmission, especially for salient morphological (a lot of the eukaryotic genome may be shaped by intragenomic dynamics though). This was at the heart of Mark Ridley’s book The Cooperative Gene: How Mendel’s Demon Explains the Evolution of Complex Beings.

Second, it probably is useful for people to review some of the science. Here is a paper which seems open access, A CRISPR-Cas9 gene drive system targeting female reproduction in the malaria mosquito vector Anopheles gambiae. 267858Anyone who is more interested in the topic needs to read Austin Burt and Robert Trivers’ Genes in Conflict: The Biology of Selfish Genetic Elements. And, as I pointed out to Antonio meiotic drive has been on the minds of population geneticists’ for decades.

Personally I haven’t given deep thought to the utilization of gene drive in public health campaigns, though it seems to be one of the favorite tools for those arguing for mosquito extermination. One the one hand the potential for ecological disturbance is not trivial, though perhaps more in a case where gene drive methods are executed badly and other species’ are inadvertently effected (my understanding is that conventional mosquito eradication has not led to ecosystem collapse anywhere). And on the other hand there is the fact that humans constantly disturb the ecology of many organisms. Genetics, even gene drive, is not magic. At least in the sense that Arthur C. Clarke would have defined magic. A lot of the fear is just fear of the unknown.

Update: Kevin Esvelt leaves a comment I’m going to reproduce below the fold. Note the emphasized section. I’m pretty shocked:

Not only is gene drive a natural phenomenon, it is ubiquitous. I would be extremely surprised if there are more than a handful of species lacking either an active gene drive element or the broken remnants of one somewhere in their genomes. This is mostly due to transposons, as the other mechanisms are limited to certain reproduction patterns, but homing endonucleases – which operate by the same mechanism as a CRISPR gene drive – are not uncommon.

Austin’s book is a great reference to the natural phenomenon, and he was also the first to propose harnessing homing endonucleases to engineer wild populations in a landmark paper back in 2003. As it turned out, retargeting homing endonucleases is incredibly difficult, and most if not all such gene drives would not exhibit high penetrance (affect a large fraction of the natural population) because they would create and be subsequent blocked by drive-resistant alleles with mutations in the target site. This is why CRISPR is important: it allows multiple sites to be targeted in order to preclude resistance.

Anyone interested in CRISPR gene drive, its capabilities, and applicable safeguards should consider reading our original open-access paper that first detailed the technology and its implications:

http://elifesciences.org/content/3/e03401

We chose to publish before performing any experiments in the laboratory in order to set an example: public notification and discussion should precede and inform experiments when technological interventions will have shared effects. A secondary motive was to ensure that other scientists used appropriate safeguards in order to avoid disaster. That disaster would not be environmental; as Razib points out, we habitually use technologies that are far more ecologically disruptive than gene drive.

Rather, the greatest risk is social. The accidental release of a synthetic gene drive element into the wild would be fairly convincing evidence that at least some scientists cannot be trusted with a technology capable of unilaterally altering shared ecosystems. Given the ubiquity of CRISPR and the simplicity of design, there’s little chance of restricting anyone who already has the requisite transgenesis capabilities from pursuing it, but the resulting loss of trust in scientists would almost certainly preclude opportunities to use gene drive interventions to benefit humanity and/or the environment.

Of course, Murphy’s Law applies: the first working example in an insect was done by a group that wanted to make homozygous fruit fly knockouts in a single step. They had not read any of our work – or Austin’s, for that matter – and were apparently unaware of natural gene drive systems at the time they ran the experiments. We may have only escaped an accidental release because the first author suggested that they do the experiments in a mosquito facility for added containment. They subsequently joined us and many others in publishing consensus recommendations for appropriate safeguards, but the same thing could happen again.

The best path towards winning public acceptance is to consciously depart from traditional research and technology development practices. Public notification and discussion should precede and inform experiments, detail safeguards to be used in advance, invite and respond to public criticism, focus on applications with clear benefits to ordinary citizens, and ensure that development is led by nonprofits (academia, government, philanthropy) to minimize perceptions of bias and conflicts of interest. GMOs are a mess because they did none of these.

A more responsive approach to science would not only improve the likelihood of acceptance and be far more ethical than the conventional development cycle, it also reduces the risk that something will go wrong simply because you have far more eyes looking at the problem. I am particularly keen to have anti-biotech activists scrutinizing my proposals, as they have the most incentive to identify flaws.

Finally, it’s worth noting that CRISPR gene drive is a very poor bioweapon. Humans can’t be directly affected because of our generation time; crops (at least in the developed world) are resistant because we use seed farms that tightly control genetics. Weaponizing fast-reproducing sexual species is not a trivial engineering problem.

But more importantly, drive elements are readily detected by sequencing, spread slowly, and can be blocked by an appropriate immunizing reversal drive element that is trivial to construct if you’ve already seen a working example of a gene drive in that species. This does not mean that drive elements could not do harm before detection and reversal – assuming we actively monitor the environment for them – only that they are comparatively poor bioweapons. I am honestly not sure what we would have done in terms of the original public disclosure were it otherwise, but it does suggest that we need a better system for advising scientists who discover potentially dangerous technologies, because there are many more boxes waiting to be opened.

Note: Some molecular population geneticists have expressed rather sanguine attitudes toward the “threat” of gene drive.

Posted in Uncategorized

Comments are closed.