Predicting the invasiveness of threshold-dependent gene drives

Gene drives hold great promise for controlling disease vectors or invasive species due to their capacity to rapidly spread through a population from a small initial release. This same property also raises serious concerns about unintended spillover into non-target populations. Threshold-dependent gene drive systems, which can spread only when introduced above a critical population frequency, have been proposed as a more controllable alternative, yet their invasion dynamics in spatially structured populations remain poorly understood. Here, we analyze invasion criteria for threshold-dependent gene drives in continuous-space populations using deterministic reaction-diffusion models and individual-based simulations that better capture the stochasticity of real-world populations. We find substantial variability in invasion outcomes in the individual-based models. Low-threshold modification drives with small fitness costs frequently spread across a wide range of release sizes, including introductions far below those required to succeed in diffusion models. In contrast, threshold-dependent suppression drives exhibit qualitatively different behavior: stochastic effects at low density can often disrupt wavefronts or produce persistent chasing cycles, generally reducing invasion success relative to diffusion-model expectations. Overall, our results show that the spatial containment of threshold-dependent gene drives is more complex than predicted by non-spatial or purely deterministic models, highlighting the importance of spatially explicit analyses when evaluating their real-world performance.