Code for resident vs. challenger LV competition model for S. pneumoniae
As well as python3, the following python modules are needed to run this code (shown are the versions we used):
- dill==0.2.7.1
- numpy==1.13.3
- matplotlib==2.1.0
- lmfit==0.9.7
- statsmodels==0.8.0
- scipy=0.19.1
- numba==0.36.2
- sdeint=0.2.1
- elfi==0.7
- networkx==1.11 (note, must be < v2)
See growth_fit.ipynb or run:
python growth_fit.py
Some example solutions and plots can be run with:
python ode_int.py --mode ode --t_com 3.8 --t_chal 24 --C_size 10
python ode_int.py --mode ctmc --t_chal 10 --C_size 20000 --beta 0.1
python ode_int.py --mode sde --t_com 3.8 --t_chal 6 --C_size 10 --beta 1.48 --R_size 10 --t_end 36 --g-RC 0.01 --g-CR 0.5 --resolution 2000
Will run the ODE, CTMC and SDE forumulations respectively, exhibiting different parameter values in each case.
Running the above, but replacing ode_int.py with time_methods.py will print the time taken to do 100 integrals to STDOUT. For example:
python time_methods.py --mode ctmc --t_chal 10 --C_size 20000 --beta 0.1
4.465044894997845
First of all use BOLFI to fit parameters based on the SDE model (takes around 6hrs):
python elfi_simulator.py --experiments experimental_data.txt
Once BOLFI has approximated the posterior, statistics about the fit, and samples from the posterior can be taken by running:
python elfi_post_sample.py --bolfi bolfi.pkl
First, use grid_model.py to run the model over a range of:
- lag in arrival time and challenger inoculum size for isogenic challengers
- competition values for intergenic challengers
# isogenic
python grid_model.py --mode ctmc --repeats 20 --output ctmc --isogenic --grid-resolution 100
# intergenic
python grid_model.py --mode ctmc --repeats 20 --threshold 10 --output ctmc.t_chal1 --intergenic --t-chal 1 --grid-resolution 100
python grid_model.py --mode ctmc --repeats 20 --threshold 10 --output ctmc.t_chal2 --intergenic --t-chal 2 --grid-resolution 100
python grid_model.py --mode ctmc --repeats 20 --threshold 10 --output ctmc.t_chal4 --intergenic --t-chal 4 --grid-resolution 100
python grid_model.py --mode ctmc --repeats 20 --threshold 10 --output ctmc.t_chal6 --intergenic --t-chal 6 --grid-resolution 100
python grid_model.py --mode ctmc --repeats 20 --threshold 10 --output ctmc.t_chal24 --intergenic --t-chal 24 --grid-resolution 100
Then use these outputs to create contour plots. First for the 'resident wins' boundary in the isogenic case:
python isogenic_contour_plot.py --runs ctmc.isogenic_runs.txt --output isogenic_ode --boundary 1 --smooth
The domains up to 24hrs separately for each intergenic case:
for i in 1 2 4 6 24;
do python intergenic_contour_plot.py --runs ctmc.t_chal$i.intergenic_runs.txt \\
--output intergenic_tchal$i --smooth --title "Intergenic challenger (t_chal = $i hrs)" \\
--stochastic;
done
The domains for 1-6hrs on one plot:
cat ctmc.t_chal*.intergenic_runs.txt > ctmc.all.intergenic_runs.txt
python intergenic_multiple_contour_plot.py --runs ctmc.all.intergenic_runs.txt --smooth