Assessing the risk of gene flow from genetically modified trees carrying mitigation transgenes

Abstract

The increasing use of genetically modified (GM) plants has raised concerns about the escape of transgenes to conventional populations. To counteract possible fitness advantages of GM plants, an advantageous ‘primary transgene’ may be linked to a ‘mitigation’ transgene that is selectively disfavored. A risk related to this technique is the possible break-up of the transgenes. This may lead to the establishment of genotypes that only express the primary transgenic trait and are therefore selectively favored over both conventional and GM genotypes. We study here how the establishment of break-up genotypes in tree populations depends on life history, pollen dispersal kernels, and the linkage between transgenes. To this end, we model the dynamics of an initially homozygous GM population and a conventional population growing next to each other. The simulations are based on a spatially explicit simulation model for population-genotype dynamics. Pollen and seed dispersal are described by fat-tailed dispersal kernels. Break-up of transgenes is due subsequent crossing of GM and conventional genotypes. The spatial pattern of pollen and seed dispersal causes break-up and establishment of new invasive genotypes to occur mostly in the conventional population. The amount of individuals of these genotypes established in the conventional population depends on the interaction between life-history traits, pollen dispersal distances and the linkage of the transgenes. Thus, risk assessment of transgene break-up requires information on the genetics, local dynamics, and dispersal of GM and conventional varieties. The approach presented here provides a tool for combining these different sources of information, and it contributes to the formulation of general rules for the management of GM tree populations.

Appendix

Table A1 Simulation parameters for the landscape and for two life-history scenarios