Gene drives do not always increase in frequency: from genetic models to risk assessment

Homing genes encode endonucleases that make a double stranded break in the DNA, destroying a target site on the homologous chromosome. When the cell repairs the break the homing allele is copied, converting a heterozygote into a homozygote. This results in gene drive (GD), an overrepresentation of the homing allele in the next generation. GD may propel CRISPR-Cas9 genes and new genes physically coupled to the GD through natural populations. I revisit the population genetic models of GD with the aim of making these models more understandable to non-specialists. What can we learn about risk evaluation from the models? A GD with no or a small effect on fitness (viability) always spreads in the population and goes to fixation. That is provided that no resistance mechanism evolves, for instance due to a mutation in the target site. However, when GDs have a large negative effect on fitness, their spread depends on a threshold or they may not spread at all. The chance of GDs increasing until fixation is much higher in systems with meiotic drive than in systems with embryo conversion. The presence or absence of a meiotic promoter is therefore relevant to take into account in the environmental risk assessment.