Modelling landscape-level approaches for mitigating the evolution of pesticide resistance

Food security is a global challenge reflected across two United Nations Sustainable Development Goals (SDGs 2 & 3). One critical component to meeting this challenge is the management of agricultural pests. Successful pest management is regularly thwarted by the rapid evolution of pest resistance to pesticides [1, 2]. Consequently, pesticide resistance is a widespread problem despite the ongoing development of various strategies to disrupt or weaken it. Such strategies often manage when and where pesticides are applied to ensure that some susceptibility to pesticides persists in the pest population [3, 4]. Testing different pesticide application strategies can be time-consuming and expensive in real agricultural systems, and such tests are constrained in their ability to control for pest characteristics such as genome properties and trait distributions. Instead, agent-based models (ABMs) can be used to simulate the social-ecological dynamics of pesticide application regimes and pest eco-evolutionary responses to determine the most effect strategies for overcoming pesticide resistance evolution [3, 5]. This project will use ABMs to develop theory on pesticide resistance management across a range of pesticide application regimes, pest life histories, and pest genome properties.

Pest management strategies are often developed on the premise that uniform and constant pesticide application will cause sustained directional selection for pesticide resistance. Hence, by relaxing selection using designated refuge areas where no pesticide is applied, alleles conferring susceptibility to pesticides can be maintained in the pest population [1, 3]. Alternatively, a combination of different types of pesticides applied in combination, or at different times and in different locations, might also be effective for maintaining susceptibility to any single pesticide [4]. Such approaches with refuges or multiple pesticides might be especially advantageous if there are trade-offs in fitness related pest traits [6, 7]. For example, if pesticide resistance incurs a cost to pest reproduction or longevity. Consequently, the efficacy of management strategies can depend on the complex relationships among pest life histories [4], mating systems [8], and genetics [9]. This project will investigate management strategies given these relationships using ABM computer simulations in the newly developed resevol R package [5, https://bradduthie.github.io/resevol/].

The resevol R package [5] is a highly flexible tool for simulating ABMs of pest evolutionary and ecological dynamics under different regimes of crop and pesticide use. Simulations are run on spatially explicit landscapes with a potentially high level of terrain detail and pre-specified farm locations. Crops and pesticides can be rotated over time for each farm independently, which can simulate different management strategies applied across the full agricultural landscape. Pests are modelled to have evolving traits that have a pre-specified covariance structure and means. These traits can affect multiple aspects of life history including movement, reproduction, metabolism, feeding ability on specific crops, and resistance to specific pesticides. Pest traits are polygenic, quantitative traits produced from explicit and unique agent genomes that have an arbitrary number of loci in which alleles undergo mutation, recombination, selection, and drift. This complexity allows simulations to fully track all aspects of pest evolution and ecology. The PhD student will have the opportunity to apply the resevol R package and develop their own model or case study to evaluate different crop and pesticide application regimes for different agricultural pest systems and provide policy recommendations for pest resistance management.

Click on an image to expand

Image Captions

Logo for the resevol R package.


The primary aim of the project is to develop new theory on pest management through the use of agent-based models (ABMs). This new theory will integrate knowledge from the literature and ABM simulations to predict the general consequences of different pest management strategies under different evolutionary, ecological, and environmental conditions. The project is divided into three initial stages presented below. These stages are flexible and might be later tailored to meet the specific interests of the student.

The first stage of the project focuses on the landscape. The student will construct an ABM with the resevol R package and run simulations that vary in landscape structure (e.g., farm locations and dimensions) and crop and pesticide rotation (e.g., random or sequential rotation of crops and pesticides). Using these simulations, the student will test how agricultural production, pest population density, and pest evolution respond to different landscape characteristics and regimes of crop and pesticide application.

The second stage of the project focuses on pest reproduction. The student will simulate pest populations with different types of reproductive systems (e.g., asexual, monoecious, or dioecious) to determine how reproductive system affects agricultural production and pest populations and which pest management regimes are most effective for each system.

The third stage of the project focuses on pest genomes. The student will investigate the effect of pest genome properties on pesticide resistance evolution. Genome properties will include loci number and effect size, recombination rate, mutation rate, and genetic constraints on pest traits. The student will use the resevol R package to vary these genome properties and determine how pest genomes modulate the efficacy of different pest management strategies.

The remainder of the project will be flexible and targeted to the student’s own research interests. This might include further development of the resevol R package with the student’s own code and simulations, application of simulations to a specific case study (e.g., the supervisory team is currently working with the agricultural pest species Helicoverpa armigera), or the development of a new model from scratch.

Project Timeline

Year 1

Review of the literature for the evolution of pesticide resistance theory, modelling, and case studies. Modelling skills development including R, C, git, and GitHub. Initial model design, parameterisation, and simulations using the resevol R package for the first stage of the project.

Year 2

Simulation of ABMs for the second and third stages of the project using the resevol R package to test agricultural production, pest population dynamics, and pesticide resistance evolution as a consequence of different pest reproductive systems and genomes.

Year 3

Further development of the resevol R package and new simulation scenarios in resevol, simulations of a specific case study of interest, or new model development and use.

Year 3.5

Thesis write up and publishing peer-reviewed papers.

& Skills

During this project, the student will join the development team of the resevol R package and learn how to develop social-ecological theory using agent-based models (ABMs). They will work with the primary supervisor to identify important theoretical questions, then design and simulate appropriate ABMs to address those questions using state of the art modelling software. The student will thereby develop expertise in model development, use, and analysis. All coding will be done collaboratively with the lead supervisor using GitHub, and software developed will be published open access.

As part of this project, the student will develop expertise in at least two programming languages (R and C). They will also learn key skills for modelling and software development, including version control (git), software documentation, R package development, and programming skills for agent-based modelling. The student will have the opportunity to write their own models from scratch or apply modelling software to a case study of their own interest.

The student will present their research at national and international conferences, workshops, and University of Stirling student symposia and seminars. They will also participate in the University of Stirling’s coding club. Finally, the student will be encouraged to publish their modelling results in international peer-reviewed journals.

References & further reading

[1] Gould, F., Brown, Z. S., and Kuzma, J. (2018). Wicked evolution: Can we address the sociobiological dilemma of pesticide resistance? Science, 360(6390):728–732.[2] Lykogianni, M., Bempelou, E., Karamaouna, F., and Aliferis, K. A. (2021). Do pesticides promote or hinder sustainability in agriculture? The challenge of sustainable use of pesticides in modern agriculture. Science of the Total Environment, 795:148625.[3] Saikai, Y., Hurley, T. M., and Mitchell, P. D. (2021). An agent-based model of insect resistance management and mitigation for Bt maize: a social science perspective. Pest Management Science, 77(1):273–284.[4] Sudo, M., Takahashi, D., Andow, D. A., Suzuki, Y., and Yamanaka, T. (2018). Optimal management strategy of insecticide resistance under various insect life histories: Heterogeneous timing of selection and interpatch dispersal. Evolutionary Applications, 11(2):271–283.[5] Duthie, A. B., Mangan, R., McKeon, C. R., Tinsley, M. C., and Bussière, L. F. (2022). resevol: an R package for spatially explicit models of pesticide resistance given evolving pest genomes. bioRxiv (preprint). https://www.biorxiv.org/content/10.1101/2022.08.22.504740v1[6] Raymond, B., Sayyed, A. H., and Wright, D. J. (2005). Genes and environment interact to determine the fitness costs of resistance to Bacillus thuringiensis. Proceedings of the Royal Society B: Biological Sciences, 272(1571), 1519–1524. https://doi.org/10.1098/rspb.2005.3103[7] Jensen, K., Ko, A. E., Schal, C., and Silverman, J. (2016). Insecticide resistance and nutrition interactively shape life-history parameters in German cockroaches. Scientific Reports, 6(June), 1–7. https://doi.org/10.1038/srep28731[8] Jacomb, F., Marsh, J., and Holman, L. (2016). Sexual selection expedites the evolution of pesticide resistance. Evolution, 70(12), 2746–2751. https://doi.org/10.1111/evo.13074[9] Ffrench-Constant, R. H. (2013). The molecular genetics of insecticide resistance. Genetics, 194(4):807–815.

Apply Now