This is an orderly repository which contains the analysis to our preprint
Non-pharmaceutical interventions, vaccination and the Delta variant: epidemiological insights from modelling England's COVID-19 roadmap out of lockdown
A sequence of tasks needs to be run with a set of parameters to generate the final results. This is sketched out in the run.R
script, though this is provided only as a form of documentation. In practice these were run over several days on a HPC.
- regions:
c("north_west", "north_east_and_yorkshire", "midlands", "east_of_england", "london", "south_west", "south_east")
- assumptions:
c("central", "optimistic", "pessimistic")
- Run the
vaccine_fits_regional
task with each region and assumption level (21 fits, each about 5 hours) - Run the
vaccine_fits_combined
task with each assumption level (3, collecting all regions) - Run the
vaccine_restart_fits_regional
task with each region and assumption level (21 fits, each about 10 hours) - Run the
vaccine_restart_fits_combined
task with each assumption level (3, collecting all regions) - Run the
vaccine_simulation
task with each assuption level, with arestart_date
ofmarch
,june
andjuly
withsensitivity=FALSE
, and also withsensitivity=TRUE
forrestart_date="july"
. The full runs usen_par = 200
but this can be reduced at the cost of more noise. Because of the large number of scenarios used, these were run across a set of 32 core nodes (up to 10 at a time). - Run the
vaccine_simulation_plots
task - Run the
vaccine_simulation_plots_sens
task
Running even the short run, as in run.R
will take ~7 hours of CPU time, less in wall time if you have more cores available.
The core requirement is our sircovid package and its dependencies. Because that package is in constant development you will probably want to pin your versions of the software to the versions we used for preparation:
remotes::install_github(c(
"mrc-ide/[email protected]",
"mrc-ide/[email protected]",
"mrc-ide/[email protected]",
"mrc-ide/[email protected]"))
However, you can always install the versions that we are using with
drat:::add("ncov-ic")
install.packages(c("sircovid", "spimalot"))
You will also need a recent orderly which can be installed with
drat:::add("vimc")
install.packages("orderly")
To install all required packages, you can use remotes:
remotes::install_deps()
MIT © Imperial College of Science, Technology and Medicine