Ecological engineering and gene drives

A gene drive is a genetic element that is inherited by more than half the offspring of a heterozygous parent. That single deviation from Mendelian segregation is the whole idea: an allele that spreads even when it lowers the fitness of the animal carrying it, and therefore spreads through a wild population from a small release rather than needing to be delivered to every individual.

Homing: the drive copies itself

The engineered version borrows a mechanism found in natural homing endonuclease genes. The construct encodes a nuclease and a guide targeting the exact chromosomal position the construct itself occupies, so it sits in a site it can cut only on the homologous chromosome — the wild-type copy. In the germline of a heterozygote, the nuclease cuts that wild-type allele; the cell repairs the break by homologous recombination using the intact chromosome as template, and that template contains the drive. The heterozygote becomes a homozygote before making gametes, and inheritance approaches one hundred per cent.

Two payload strategies follow. A suppression drive targets a gene required for female fertility or viability, so the population collapses as the drive fixes. A modification drive carries a cargo — for example an effector blocking parasite development in the mosquito — leaving population size intact and changing what the population can transmit.

Resistance is generated by the mechanism itself

This is the field’s central technical fact. Homing depends on the break being repaired by homologous recombination, but the same break in the same cell can instead be closed by non-homologous end joining. That produces a small indel at the cut site — an allele that is no longer recognised by the guide, is not converted, and if it preserves gene function is favoured by selection in a suppression drive. One such allele arising anywhere in a large population can be enough to rescue it. This is not an off-target problem; it is the on-target repair competition described on the genome editing page, acting at population scale.

The engineering responses are honest about it. Multiplexed guides require several independent resistant mutations at once. Targeting a sequence so functionally constrained that almost any indel is itself lethal — a conserved region of a female fertility gene is the worked example — makes resistant alleles that survive selection rare rather than merely infrequent.

Designs that give back the irreversibility

The property that makes a drive powerful is also what makes it hard to authorise: released, it does not stay where it was put. Split designs separate the nuclease from the guide so the drive cannot propagate without both. Daisy-chain drives arrange elements so each drives the next and the chain exhausts itself after a bounded number of generations. Threshold-dependent designs spread only above a release frequency, so a small accidental escape is eliminated by selection rather than amplified. Each buys containment by reducing reach.

Non-genetic population interventions — irradiated sterile males, and Wolbachia releases that alter reproduction or transmission — sit alongside these and are already field-deployed at scale. This page does not cover the governance and consent frameworks that field release requires.