The genomics of insecticide resistance: insights from recent studies in African malaria vectors

Abstract

Over 80% of the world's population is at risk from arthropod-vectored diseases, and arthropod crop pests are a significant threat to food security. Insecticides are our front-line response for controlling these disease vectors and pests, and consequently the increasing prevalence of insecticide resistance is of global concern. Here we provide a brief overview of how genomics can be used to implement effective insecticide resistance management (IRM), with a focus on recent advances in the study of Anopheles gambiae, the major vector of malaria in Africa. These advances unlock the potential for a predictive form of IRM, allowing tractable feedback for stakeholders, where the latest field data and well parameterised models can maximise the lifetime and effectiveness of available insecticides.

Figure 1

Insecticide resistance management flow diagram. Reactive IRM — an example IRM work flow without an active genomic component. 1. Insects are sampled from a region undergoing a vector control campaign, these samples can be subjected to a bioassay to determine their IR phenotype after which their DNA is collected. 2. Molecular assays for a small number of previously established IR associated genetic loci are conducted on the DNA to determine potential causal genotypes of the IR phenotype, to genetically characterise the population, this can help determine the mode of resistance, for example, target site/metabolic. 3. This can provide useful information about insecticide resistance, but the speed with which it can be passed to vector programme managers in a readily useable format may be delayed by processing time and by the fact that these assays are often conducted outside the country of collection. 4. Input from molecular assays can be used to improve IRM, but two key limitations mean that this approach is unlikely to be sufficient to fully prevent insecticide resistance, turnaround is too slow (months/years) and only established IR variants can be detected. Predictive IRM — an example IRM work flow with an active genomic component. 1. Insects are sampled from a region undergoing a vector control campaign, representative ‘sentinel’ sites within the region are sampled repeatedly over time. Ideally the initial time points are taken before the IRM is rolled out. These samples can be subjected to a bioassay to determine their IR phenotype after which DNA is collected. 2. DNA is sequenced in the country of collection and this, in concert, with advances in sequencing speed, reduces the time taken to generate data. 3. Data produced from longitudinal sampling and whole genome sequencing can be used to parameterise predictive IR models and GWAS/GWSS can locate novel IR loci allowing molecular assays to cover all potential IR linked variants/haplotypes in natural populations. 4. With strategic whole genome sequencing to update assays and model parameters, most samples can then skip the WGS step and be quickly and cheaply assayed for IR linked loci in-country with minimal technological requirements. 5. As all data is already within the country of collection, it can be quickly passed to national/local vector control program managers. 6. Using input from predictive models and genotype/phenotype associations, IRM can be modified rapidly enough (weeks/months) to avert the emergence and spread of insecticide resistance. The effectiveness of these modifications is monitored as the cycle begins again, and further improvements to IRM can be made as needed.