Skip to content

mpievolbio-scicomp/cancer_sim

Repository files navigation

Build Status Documentation Status Binder DOI

Documentation

Documentation for CancerSim, including this README and the API reference manual is hosted on readthedocs.

Background

Cancer is a group of complex diseases characterized by excessive cell proliferation, invasion, and destruction of the surrounding tissue  [1] . Its high division and mutation rates lead to excessive intratumour genetic heterogeneity which makes cancer highly adaptable to environmental pressures such as therapy  [2] . This process is known as somatic evolution of cancer. Throughout most of its existence a tumour is inaccessible to direct observation and experimental evaluation. Therefore, computational modelling can be useful to study many aspects of cancer. Some examples where theoretical models can be of great use include early carcinogenesis, as lesions are clinically observable when they already contain millions of cells, seeding of metastases, and cancer cell dormancy [3] .

Statement of Need

Advanced cancer simulation software [@waclaw2015] often exhibit a prohibitively steep learning curve especially for new students in the field of somatic evolution of cancer. A software package that is accessible, simple to use, and yet covers the essential biological processes of cancer growth is needed to provide an entry point for students and newcomers to mathematical oncology.

Here, we present CancerSim, a software that simulates somatic evolution of tumours. The software produces virtual spatial tumours with variable extent of intratumour genetic heterogeneity and realistic mutational profiles. Simulated tumours can be subjected to multi-region sampling to obtain mutation profiles that are realistic representation of the sequencing data. This makes the software useful for studying various sampling strategies in clinical cancer diagnostics. An early version of this cancer evolution model was used to simulate tumours subjected to sampling for classification of mutations based on their abundance [4].

Cancer growth model

Our model is abstract, not specific to any neoplasm type, and does not consider a variety of biological features commonly found in neoplasm such as vasculature, immune contexture, availability of nutrients, and architecture of the tumour surroundings. It most closely resembles the superficially spreading tumours like carcinoma in situ, skin cancers, or gastric cancers, but it can be used to model any tumour on this abstract level.

The tumour is simulated using a two-dimensional, on-lattice, agent-based model. The tumour lattice structure is established by a sparse matrix whose non-zero elements correspond to the individual cells. Each cell is surrounded by eight neighbouring cells (Moore neighbourhood). The value of the matrix element is an index pointing to the last mutation the cell acquired in the list of mutations which is updated in each simulation step.

Simulation parameters

The simulation advances in discrete time-steps. In each simulation step, every tumour cell in the tumour that has an unoccupied neighbour can divide with a certain probability (set by the parameter division_probability). The daughter cell resulting from a cell division inherits all mutations from the parent cell and acquires a new mutation with a given probability mutation_probability). A new mutation that changes death and birth probability of cell can be introduced at into random cell at the specific time step defined by adv_mutation_wait_time. By changing fitness parameters of a mutant cell adv_mutant_division_probability and adv_mutant_death_probability one can model various evolutionary processes like emergence of a faster dividing sub-clone or selective effects of a drug treatment.

The simulation allows the acquisition of more than one mutational event per cell (number_of_mutations_per_division). In that case, variable amounts of sequencing noise [6] can be added to make the output data more biologically realistic. The key parameters number_of_generations, division_probability and death_probability determine the final size of the tumour, while the degree of intratumour heterogeneity can be varied by changing the mutation_probability parameter. For neutral tumour evolution, parameter adv_mutant_division_probability and adv_mutant_death_probability must be the same as division_probability and death_probability.

Throughout the cancer growth phase, CancerSim stores information about the parent cell and a designation of newly acquired mutations for every cell. Complete mutational profiles of cells are reconstructed a posteriori based on the stored lineage information.

The division rules which allow only cells with empty neighbouring nodes to divide, cause exclusively peripheral growth and complete absence of dynamics in the tumour centre. To allow for variable degree of growth inside the tumour, we introduced a death process. At every time step, after all cells attempt their division, a number of random cells die according to death_probability and adv_mutant_death_probability and yield their position to host a new cancer cell in a subsequent time step.

Simulation results

After the simulation, the tumour matrix, and the lists of lineages and frequencies of each mutation in the tumour are exported to files. Furthermore, the virtual tumour can be sampled and a histogram over the frequency of mutations will be visualised. Alternatively, a saved tumour can be loaded from file and then be subjected to the sampling process.

Installation

CancerSim is written in Python (version >3.5). We recommend to install it directly from the source code. To download the code:

EITHER clone the repository:

$> git clone https://github.com/mpievolbio-scicomp/cancer_sim.git

OR download the source code archive:

$> wget https://github.com/mpievolbio-scicomp/cancer_sim/archive/master.zip
$> unzip master.zip -d cancer_sim

Change into the source code directory

$> cd cancer_sim

We provide for two alternatives to install the software after it was downloaded:

Alternative 1: conda

New conda environment

We provide an environment.yml to be consumed by conda. To create a fully self-contained conda environment (named casim):

$> conda env create -n casim --file environment.yml

This will also install the cancer simulation code into the new environment.

To activate the new conda environment:

$> source activate casim

or

$> conda activate casim

if you have set up conda appropriately.

Install into existing and activated conda environment

To install the software into an already existing environment:

$> conda activate <name_of_existing_conda_environment>
$> conda env update --file environment.yml

Alternative 2: pip

The file requirements.txt is meant to be consumed by pip:

$> pip install -r requirements.txt [--user]

The option --user is needed to install without admin privileges.

Installed module

After installation, CancerSim is available in python as the casim module. E.g. in a python script, one would import the module as:

>>> from casim import casim

Testing

Although not strictly required, we recommend to run the test suite after installation. Simply execute the run_tests.sh shell script:

$> ./run_tests.sh

This will generate a test log named casim_test@<timestamp>.log with <timestamp> being the date and time when the test was run. You should see an OK at the bottom of the log. If instead errors or failures are reported, something is wrong with the installation or the code itself. Feel free to open a github issue at https://github.com/mpievolbio-scicomp/cancer_sim/issues and attach the test log plus any information that may be useful to reproduce the error (version hash, computer hardware, operating system, python version, a dump of conda env export if applicable).

The test suite is automatically run after each commit to the code base. Results are published on travis-ci.org.

High-level functionality

Setting the simulation parameters

The parameters of the cancer simulation are specified in a python module or programmatically via the CancerSimulationParameters class. A documented example params.py is included in the source code (under casim/params.py) and reproduced here:

$> cat casim/params.py
################################################################################
#                                                                              #
# Commented casim parameter input file.                                        #
# Valid settings are indicated in parentheses at the end of each comment line. #
# [0,1] stands for the closed interval from 0 to 1, including the limits; ||   #
# means "or".                                                                  #
#                                                                              #
################################################################################

# Number of mesh points in each dimension (>0)
matrix_size = 1000

# Number of generations to simulate (>0).
number_of_generations = 20

# Probability of cell division per generation ([0,1]).
division_probability = 1

# Probability of division for cells with advantageous mutation ([0,1]).
adv_mutant_division_probability = 1

# Fraction of cells that die per generation ([0,1]).
death_probability = 0.1

# Fraction of cells with advantageous mutation that die per generation ([0,1]).
adv_mutant_death_probability = 0.0

# Probability of mutations ([0,1]).
mutation_probability = 1

# Mutation probability for the adv. cells ([0,1]).
adv_mutant_mutation_probability = 1

# Number of mutations per cell division (>=0).
number_of_mutations_per_division = 10

# Number of generations after which the beneficial mutation occurs (>=1).
adv_mutation_wait_time = 10

# Number of mutations present in first cancer cell (>=0).
number_of_initial_mutations = 150

# Tumour multiplicity ("single" || "double").
tumour_multiplicity = "single"

# Sequencing read depth (read length * number of reads / genome length).
read_depth = 100

# Fraction of cells to be sampled ([0,1]).
sampling_fraction = 0.1
    
# Sampling position (list of (x,y) coordinates in the range [0,matrix_size-1]).
# If left blank or None, random position will be chosen.
# sampling_positions = None # This will randomly set a single sampling position.
sampling_positions = [(500,500),(490,490)]

# Plot the tumour growth curve (True || False).
plot_tumour_growth = True
    
# Export the tumour growth data to file (True || False).
export_tumour = True

Here, we simulate a single 2D tumour on a 1000x1000 grid (matrix_size=1000) for a total of 20 generations (number_of_generations=20). On average, both healthy and mutant cells divide once per generation (division_probability). The first cancer cell carries 150 mutations (number_of_initial_mutations=150); both healthy and mutant cells aquire 10 new mutations (number_of_mutations_per_division=10) in each generation with a certainty of 100% (mutation_probability=1). The advantageous mutation happens in the 10th generation (adv_mutation_wait_time=10). Mutant cells with advantageous mutations live on forever (adv_mutant_death_probability=0) while healthy cells die with a rate of 0.1 per generation (death_probability=0.1).

After completion of the last generation, two spatial samples are taken, one from the tumour center and one from a slightly more lateral position (sampling_positions = [(500,500),(490,490)]). Each sample contains 10% (with respect to the whole tumour size) closely positioned tumour cells (sampling_fraction=0.1). The samples are subject to genetic sequencing with a read depth of 100 (read_depth=100). The data is written to disk (export_tumour=True) and plots showing the mutation histograms for the whole tumour as well as for the sampled part of the tumour are generated. Furthermore, a plot showing the tumour growth over time is saved (plot_tumour_growth=True).

Users should start with the template and adjust the parameters as needed for their application by setting experimentally or theoretically known values or by calibrating the simulation output against experiments or other models.

Run the example

The simulation is started from the command line. The syntax is

$> python -m casim.casim [-h] [-s SEED] [-p PARAMS] [-o DIR]

SEED is the random seed. Using the same seed in two simulation runs with identical parameters results in identical results. If not given, SEED defaults to 1. PARAMS should point to a python parameter file. If not given, it defaults to params.py in the current working directory. If that file does not exist, default parameters are assumed. DIR specifies the directory where to store the simulation log and output data. If not given, output will be stored in the directory casim_out in the current directory. For each seed, a subdirectory cancer_SEED will be created. If that subdirectory already exists because an earlier run used the same seed, the run will abort. This is a safety catch to avoid overwriting data from previous runs.

Output

After the run has finished, you should find the results in the specified output directory:

$> ls out/cancer_1/simOutput  
growthCurve.pdf  mut_container.p              sample_out_490_490.txt
mtx.p            sampleHistogram_490_490.pdf  sample_out_500_500.txt
mtx_VAF.txt      sampleHistogram_500_500.pdf  wholeTumourVAFHistogram.pdf   

Let's take a look at the .txt files. They contain the simulation output: mtx_VAF.txt is a datafile with three columns: mutation_id lists the index of each primary mutation, additional_mut_id indexes the subsequent mutations that occur in a cell of a given mutation_id; frequency is the frequency which at a given mutation occurs.

Corresponding to the given sample positions, there is one sample_out_XXX_YYY.txt and one sampleHistogram_XXX_YYY.pdf for each position. The .txt filel lists all mutations of the artificial sample taken from the whole tumour. Columns are identical to mtx_VAF.txt.

The .pdf files are plots of sampled tumour histogram. wholeTumourVAFHistogram.pdf is the histogram for the complete tumour. You should see figures similar to these:

Whole tumour histogram:
Whole tumour histogram

Central sample histogram:
Sampled tumour histogram

Lateral sample histogram:
Sampled tumour histogram

The remaining output files are serialized versions ("pickles") of the tumour geometry as a 2D matrix (mtx.p) and the mutation list (list of tuples listing the cancer cell index and the mutation ID of each tumour cell, mut_container.p).

Example notebooks

An example demonstrating how to parametrize the simulation through the CancerSimulationParameters API is provided in the accompanying jupyter notebook at quickstart_example.ipynb`. Launch it in binder.

In run_dump_reload_continue.ipynb, we demonstrate how to use the restart capability to modifiy tumour growth parameters in the middle of a run. In this way, one can model different phases of tumour growth, e.g. tumour dormancy or onset of cancer therapy.

Community Guidelines

As an Open Source project, we welcome all contributions to CancerSim. We recommend the usual github workflow: Fork this repository, commit your changes and additions to the fork and create a pull request back to the master branch on this repository. If uncertain about anything, please create an issue at https://github.com/mpievolbio-scicomp/cancer_sim/issues.

Comments, bug reports, or other issues as well as requests for support should be submitted as a github issue. Please check the list of issues if your problem has already been addressed. We will do our best to respond in a timely manner.

References

[1] J. C. A. Vinay Kumar Abul K. Abbas, Robbins Basic Pathology, 10th ed. (Elsevier, 2017). ISBN: 9780323353175.

[2] S. Turajlic, A. Sottoriva, T. Graham, and C. Swanton, Nat Rev Genet (2019). DOI: 10.1038/s41576-019-0114-6

[3] P. M. Altrock, L. L. Liu, and F. Michor, Nat Rev Cancer 15, 730 (2015). DOI: 10.1038/nrc4029

[4] L. Opasic, D. Zhou, B. Werner, D. Dingli, and A. Traulsen, BMC Cancer 19, 403 (2019). DOI: 10.1186/s12885-019-5597-1

[5] B. Waclaw, I. Bozic, M. E. Pittman, R. H. Hruban, B. Vogelstein, and M. A. Nowak, Nature 525, 261 (2015). DOI: 10.1038/nature14971

[6] M. J. Williams, B. Werner, C. P. Barnes, T. A. Graham, and A. Sottoriva, Nature Genetics 48, 238 (2016). DOI: 10.1038/ng.3489