Spatiotemporal modelling of the spread of insecticide resistance in Africa

This repo contains models and code to model the evolution and spread of insecticide resistance among malaria vectors in Africa. It uses a semi-mechanistic model to predict future levels of resistance across the continent, fitting closely to resistance data.

The published maps and time series are of the two-stage model: the dynamical model, corrected by a second-stage geostatistical model of its residuals (R/two_stage_maps.R, doc/two_stage_plan.md). The cross-validation figures and the dynamical model’s own parameters (e.g. the covariate effects) show the dynamical model separately.

Running order of scripts:

To be tidied: this may not be the correct order

1. packages.R
2. functions.R
3. prep_admin.R
4. prep_country_borders.R
5. prep_bioassays.R
6. prep_rasters.R
7. prep_net_use_pyrethroid.R
8. calculate_ingredient_fractions.R
9. fit_model_glm.R
10. fit_model.R
11. chain_mode_check.R, then drop_stuck_chains.R with DROP_CHAINS from its .drop file: leaves out chains in the minor mortality-floor mode (#37)
12. illustrate_validation.R
13. mtm_ir_explore.R
14. ploidy_demo.R
15. predict.R
16. two_stage_maps.R: prepare, fit 1-9, map for the six types not in LLINs and llin_effective, then figures (doc/two_stage_plan.md). The figure scripts below that show predictions read its outputs
17. summarise_model_fit.R
18. visualise_colony_net_bioassay.R
19. visualise_data.R
20. fig_admin_maps.R
21. fig_baseline_susceptibility.R
22. fig_bioassay_maps.R
23. fig_covariate_effects.R
24. fig_covariate_maps.R
25. fig_data_distribution.R
26. fig_illustrate_bioassay_variability.R
27. fig_internal_validation.R
28. fig_ir_maps.R
29. fig_ir_map_2025.R
30. fig_ir_map_2000_2030.R
31. fig_ento_epi_impact.R
32. fig_temporal_preds_data.R
33. fig_temporal_preds_net_use.R