Rescue by gene swamping as a gene drive deployment strategy

Abstract

Gene drives are genetic constructs that can spread deleterious alleles with potential application to population suppression of harmful species. Given that a gene drive can potentially spill over to other populations or even other species, control measures and fail-safes strategies must be considered. Gene drives are designed to generate a rapid demographic decline, while at the same time generating a dynamic change in the population's genetics. Since these evolutionary and demographic processes are linked and are expected to occur at a similar time-scale during gene drive spread, feedback between these processes may significantly affect the outcome of deployment. To study this feedback and to understand how it affects gene drive spillovers, we developed a gene drive model that combines evolutionary and demographic dynamics in a two-population setting. The model demonstrates how feedback between evolutionary and demographic dynamics can generate additional outcomes to those generated by the evolutionary dynamics alone. We identify an outcome of particular interest, where the short-term suppression of the target population is followed by gene swamping and loss of the gene drive. This outcome could be useful for designing gene drive deployments that temporarily suppress the population, but ultimately do not remain in the population. Using our model, we demonstrate the robustness of this outcome to spillover and to the evolution of resistance, and suggest that it could be used as a fail-safe strategy for gene drive deployment.

Incorporating ecology into gene drive modelling

“…Therefore, modelling and understanding the behaviour of gene drives in a spatial context is crucial. Current spatial models seek to understand the spatiotemporal behaviour of different gene drives under different genetic mechanisms or ecological scenarios, and thereby aim to develop potential strategies for mitigating the risks of spillover by determining conditions under which the spatial localization of the gene drive can be attained (Champer, Zhao, et al, 2020; Greenbaum et al, 2021; Harris & Greenbaum, 2022).…”

“…This behaviour is typically restricted to cases in which the migration rates between the populations are fairly low (below m ≈ 0.1 in the model in Figure 2a, but to achieve some robustness in parameter specifications typically below m ≈ 0.01; see Greenbaum et al, 2021). The simple two‐population model, therefore, already illustrates the possibility of confining a gene drive to a certain geographic location by tuning the genetic parameters and mechanism of the drive (see examples in Tanaka et al, 2017; Noble et al, 2018; Willis & Burt, 2021; Greenbaum et al, 2021; Harris & Greenbaum, 2022; Beaghton & Burt, 2022; Gamez et al, 2021; Champer, Yang, et al, 2020).…”

“…(iii) Only target population is affected (right panel). When adding demography to the model (dashed lines), other outcomes such as fluctuations, oscillations and transient suppression may arise (model following Harris and Greenbaum (2022)). The different outcomes (i)–(iii) were generated by varying the fitness cost of the gene drive.…”

“…With additional features added to this simplistic model, other types of gene drive behaviours can emerge. For example, when explicitly adding demography to the two‐population model, transient demographic bottlenecks or oscillations in population sizes and gene drive frequencies may emerge (Harris & Greenbaum, 2022) (Figure 2a). These behaviours are caused by feedback between the gene drive spread, demographic suppression caused by the gene drive and migration rates.…”

“…Higher gene flow can also increase the generation of resistance alleles: non‐homologous end‐joining (NHEJ) events can generate alleles that block the CRISPR conversion mechanism by altering the target sequence. The likelihood of these events is dependent on the frequency of heterozygotes (Unckless et al, 2017), which increases as a function of gene flow (Harris & Greenbaum, 2022). Thus, considering both topology and the level of gene flow is important for understanding how certain types of population structure could increase the probability of resistance alleles emerging.…”