OUCRU Malaria Sim v3.0.2

"OUCRU Malaria Sim" is an individual-based microsimulation of malaria transmission developed in C++. The folders MalariaCore and MalariaSimulation contain the source code and input files that are needed to compile and run the simulation. The simulation was designed to compare multiple different ways of distributing antimalarial drugs in a human population.

This version of the simulation includes transmission between hosts (without explicit implementation of mosquitoes), within-host parasite development, explicit modeling of parasite density and its effects on transmission, multi-clonal infections in hosts, pharmacokinetics, pharmacodynamics, mutation to multiple types of drug-resistant genotypes, symptomatic and asymptomatic infection, age-dependent immune acquisition, variation in host biting rates (without explicit implementation of mosquitoes), and multiple types of population-level drug deployment strategies.

This version of the simulation was used to generate the results published in

Optimal population‐level deployment of artemisinin combination therapies. Nguyen TD, Olliaro P, Dondorp AM, Baird JK, Lam HM, Farrar J, Thwaites GE, White NJ, Boni MF. Lancet Global Heath, in press, 2015.

In the individual Figure folders in this repository you will find the raw simulation outputs that were used to generate the figures in this paper.

Below, you will find instructions for compiling and running the simulation. Please contact Tran Dang Nguyen (uut47@psu.edu) or Maciej Boni (maciej.boni@ndm.ox.ac.uk) if you have any questions about the steps below.

::::::::::

COMPILING

::::::::::

The simulation will only run on Mac and Linux/Unix systems.

1. You need to have the system c++ compiler installed. We recommend the GNU g++ compiler.

2. You will need to install the BOOST libraries (www.boost.org). We used version 1.57 of the BOOST libraries for development, but there should not be any compatibility issues with future versions. Please let us know if you find any.

On Linux, if you have the apt tool insalled (Ubuntu has this by default), you can install BOOST by

sudo apt-get install libboost-dev apt-get install libboost-dev

depending on what permissions you have on your system.

On Mac, you can install the BOOST libraries with HomeBrew.

3. You will need to install the YAML markup language library

On Linux,

sudo apt-get install libyaml-cpp-dev

And on Mac use HomeBrew,

4. You will need to install the GSL library (www.gnu.org/software/gsl/). Again, for Linux:

sudo apt-get install libgsl0-dev

and use HomeBrew on Mac.

5. You should now be able to build the executable by navigating to the 'MalariaSimulation' directory with

cd MalariaSimulation

and typing

make -j 1 CONF=Release

which is a case sensitive command. You can substitute "-j 2" or "-j 3" above if you want to use more than one processor for building the project. The build should just take a few minutes.

::::::::::

RUNNING

::::::::::

6. To check if the simulation compiled correctly, type

./dist/Release/GNU-Linux-x86/malariasimulation

and this will show you a list of usage modes.

Type

./dist/Release/GNU-Linux-x86/malariasimulation -c sample_input_file.yml

to run the simulation with the sample input set of parameters provided. This runs the simulation for two years (after a burn-in stage of 11 years) on a population of 100,000 individuals. This should take between 1 and 4 minutes, depending on your hardware.

Note that many of the variables are not reported during the burn-in stage (first 4000 days) when the dynamics settle to an endemic equilibrium. The first line of the output is the random number seed, and the last line is a list of summary statistics. The remainder of the output is divided into these twenty columns.

1. time (day)

2. days of Artemisinin Monotherapy Use (AMU) - note that this can be a non-integer; see supplementary materials in 2015 LGH paper

3. Number of Treatment Failures (NTF) - discounted

4. NTF - non-discounted

5. NTF over the last 30 days - discounted

6. Clinical episodes over the last 30 days

7. Number of drug-resistance mutation events that occured in the last 30 days

8.  fraction of clonal parasite populations that have exactly one resistance mutation
   

9. fraction of clonal parasite populations that have exactly two resistance mutations

10. fraction of clonal parasite populations that have exactly three resistance mutations

11. fraction of clonal parasite populations that have exactly four resistance mutations

12. fraction of clonal parasite populations that have the artemisinin resistance mutation

13. "total resistance" - a number between zero and one that gives you the probability that a parasite population will be resistant to an individual drug, if the parasite population and the drug are chosen randomly (see supplementary appendix to 2015 LGH paper)

14. Placeholder text

15. Blood-slide prevalence, across all ages

16. Blood-slide prevalence, 2-10 age group

17. Blood-slide prevalence, 10-15 age group

18. Blood-slide prevalence, 15-45 age group

19. EIR over the past 365 days (this will be zero for the first year of the simulation)

20. Multiplicity of Infection (MOI) - this is the mean MOI over infected hosts