Code for "Minimizing cell number fluctuations in self-renewing tissues with a stem cell niche"

Rutger Kok, Sander Tans, Jeroen van Zon
Phys. Rev. E 108, 064403

This repository contains the code to reproduce the results of the above paper, as well as instructions for running your own simulations.

Workflow for reproducing results

You only need Python and numpy to run the simulations. The scripts starting with fig_ reproduce the stated figure panel(s). For example, fig_3abc_plot_two_compartment_intro.py reproduced Figure 3a-c.

Running your own simulation for certain parameters

import numpy

from stem_cell_model import tools
from stem_cell_model.parameters import SimulationParameters, SimulationConfig
from stem_cell_model.results import MultiRunStats
from stem_cell_model.two_compartment_model_space import run_simulation_niche

T = (16.153070175438597, 3.2357834505600382)
params = SimulationParameters.for_S_alpha_and_phi(S=1114, alpha_n=0.025, alpha_m=-0.275, phi=1.0, T=T, a=100/T[0])
random = numpy.random.Generator(numpy.random.MT19937(seed=1))
total_run_time = 100000

# Run the actual simulation
run_data = MultiRunStats()
while run_data.t_tot < total_run_time:
    config = SimulationConfig(t_sim=total_run_time - run_data.t_tot, random=random)
    results = run_simulation_niche(config, params)
    run_data.add_results(results)

# Collect results
stats = tools.get_single_parameter_set_statistics(run_data)

The stats object has properties like stats.n_mean. You can store simulation results using Pickle and run_data.to_dict().

Running a parameter sweep

See for example two_compartment_model_sweep_well_mixed.py for a concise example. This scripts saves the results to a folder. It uses the stem_cell_model.sweeper module, which uses all cores of your PC.

Read back the results using stem_cell_model.sweeper.load_sweep_results(folder).

Optimizing the code

See the script profile_two_compartment.py for an example of how to use the cProfile module (built-in into Python). This module profiles your code, which shows you what parts are slow, and therefore helps you optimizing the code.