Small input data files are stored directly in this Github repository in the data subfolder.
There is two larger data files (> 150MB) which are stored through the Git Large File Storage System (git lfs) and these will not download automatically with a git pull command. We recommend to download these files with your browser. In principle, you can also download them by activating the git lfs in your local git folder, but we have had bad experiences with this tool and the manual download avoids that.
For demonstration, we also include the output files generated by the code in the folder output. These will be overwritten as you run the notebooks, but we decided to include them in case you want to re-use the trait dataset directly.
This code runs with Python Python 3.7.13 and requires a list of packages. You can install them with pip -r requirements.txt.:
The Jupyter notebooks are (roughly) numbered by the sequence in which they can be executed. Notebooks with higher number tend to build on code and output created in the notebooks with lower number.
We note that some of the python scripts in this folder are actually an output of the jupyter notebooks. As you run the notebooks (e.g. 00_develop_pairwise_simulation_code.ipynb), this will generate .py file from scratch.
The figures are created and saved from the Jupyter notebooks into an extra sub-folder figures. If this subfolder is not present, you might need to create it first otherwise Python may complain.
We include a Makefile that allows to run the same Jupyter notebook with many different settings based on a Python package called papermill. For example, we use this to simulate pairwise competitions with different initial mutant abundance and input trait variation. Check out the Makefile targets for these automated recipes.