Consider the “gene drive” idea – there are a lot of variations, but the general idea is that you introduce a genetic sequence into an organism that can bias (drive) its own inheritance into the next generation. This is a thumb-on-the-scale unnatural selection if ever there was one, because that biased inheritance is outside of any fitness advantage that the new sequence might bring with it. In fact, a number of gene drive ideas have the opposite sign, conferring catastrophic unfitness in order to wipe out pathogens and disease-vector organisms.
Gene drives of various kinds show up in nature, though, when a gene has some sort of ability to control its own transmission. These are the so-called “selfish genes”, and some of these have no fitness advantage (or even some disadvantage) in the organisms themselves. There are a lot of potential mechanisms for this (see that link for a good review), but what you don’t see are the total-wipeout forms just mentioned, which is what we has humans might like to do to (say) mosquitos or tsetse flies. The advent of CRISPR-Cas9 technology has really brought a lot more attention to these ideas, because they make them far more possible, for better or worse.
The most straightforward of these would be a “homing endonuclease” gene drive. These take advantage of the double-strand-break DNA repair mechanism where a chromosome gets repaired by copying the sequence from a homologous chromosme – you end up with a stretch that’s no longer heterozygous (it’s now homozygous for whatever the donor strand sequence is), but at least it works, as opposed to being broken. A homing endonuclease gene drive (HEG) works by deliberately breaking a recipient allele at the right spot (through the expression of the Cas9 protein and a guide RNA), set up for the donor allele to copy over the whole deliberate-break-it Cas9/guide RNA machinery sequence into the broken recipient chromosome. It’s an evil trick: “Oh darn, I see that for some reason you’ve got a double-strand break, allow me to start homologous DNA repair with the appropriate template which I happen to have handy right here (inserts the new material into the recipient allele to make sure that the process happens again!) So you’re making things deliberately homozygous for this, jamming this new sequence into the next generation of cells at a much higher rate than you’d ever get naturally. And then that one does it even more.
You need to design this stuff carefully. If you induce breaks at a too-common sequence, then you’ll bollix the entire cellular reproduction process up because of DNA repair stress. You need to pick a pretty rare spot that’s as uniquely matched to your desired homologous chromosome, to keep the repairs from being made with others that match up well enough but aren’t what. you’re pushing. And if you want to get this into a significant part of the resulting population, the target will need to be a highly conserved one. But even with such constraints, this can still be done with all sorts of genetic traits.
You could, for example, introduce a gene to help with disease resistance (if you know a good one) or perhaps something that confers susceptibility to some sort of insecticide or other intervention. And you can see what happens if you start stuffing in a new sequence that confers (say) specific male or female sterility or messed-up mating behavior: a hard population crash. Unfortunately, there are a number of mechanisms that can allow the organisms to jump the fences and escape the system that you’ve prepared for them, phenomena that have been observed as various gene drive systems have been tried out in yeast, model plants like Arabidopsis, insects (fruit flies, mosquitos), and mice. For starters, you can get mutations that make your guide RNA stop guiding. At the other end of the scale, there’s also the very real possibility that your gene drive, if it works, might work all too well and spread to similar populations that you weren’t aiming for.
That’s led to a lot of proposals for “confinable” gene drives, but some of them aren’t going to be strong enough to really suppress a population, and with the ones that are, there are worries that some of them might prove to not be as confinable as planned. Another set of ideas (the toxin-antidote gene drives) has been proposed to circumvent those last problems. These target (in different schemes) genes where you disruption of one copy allows the organism to survive, but disruption of both will kill it (Toxin-Antidote Recessive Embryo, TARE) or, in the Toxin-Antidote Dominant Embryo (TADE) scheme, you target genes where both copies have to be functional for an organism to survive at all (such things do exist). In general, you set things up where you disrupt something essential and provide some sort of (tunable) rescue via the gene drive, which is also of course introducing the trait(s) that you want pushed. This new paper looks at simulations of TADE drives, which had been though to be too weak in most cases to spread well, and finds that they actually could work. It might even be able to avoid the “chasing” problem, where gene-modified individuals are dying off, but wild-type individuals move in and repopulate the empty ecological niches behind them quickly enough to render the whole process ineffective.
This stuff is getting closer to being tried out, particularly in mosquito species that are vectors for malaria. The potential for benefit is huge, but there are of course ways for things to go wrong, not all of which we can anticipate. As you’ll see from those two links, one of the big areas of research is designing controllable trials of such ideas that are large enough to reveal such problems without fully letting things rip in the wild just yet – everything from progressively larger caged enclosures all the way up to isolated offshore islands. Look for many such studies to be run, because in the end, the benefits are just too huge (elimination of terrible human diseases, control of invasive species that are messing up other ecosystems, etc.) These ideas still need a lot of tire-kicking, but in the end they’re just too useful to ignore.
