Welcome to dingo-stats, a python package for statistics on flux sampling datasets!

Installation

To run the package locally we recommend to create a virtual conda environment.

Make sure you have conda installed. To create the environment type:

conda create -n dingo-stats python=3.10.18

And then type y when this message appears : "The following NEW packages will be INSTALLED: ..."

To activate the environment type:

conda activate dingo-stats

dingo-stats relies on some functions implemented in the dingo package.

To install dingo type the following inside the activated environment:

CFLAGS="-I/usr/include/suitesparse" pip install --force-reinstall dingo-walk==0.1.6

And then install the rest dependencies of the package with:

python -m pip install --force -r requirements.txt 

Structure of the dingo-stats repository

• src: folder contains all the source code of the package
• tests: provides tutorials on how to use the functions found in src
• analysis: contains extended tutorials aiming to produce biologically significant knowledge
• ext_data: contains all external data (e.g. models, maps) used in tests

Suggested order to explore implemented functions

To better understand how to use the functions from dingo-stats you may examine the jupyter notebooks under the tests folder in the following order:

• tests/load_modify_sample.ipynb alongside tests/pathways.ipynb
• tests/loopless.ipynb
• tests/escher_maps.ipynb
• tests/distributions_comparison.ipynb
• tests/copulas.ipynb
• tests/correlations.ipynb
• tests/clustering.ipynb
• tests/graphs.ipynb

Parts of the functions are used with the same way multiple times across the notebooks. To see a complete analysis from loading a model to performing all-possible statistics check the:

• analysis/complete_workflow.ipynb

How to build the ReadTheDocs locally

cd docs
make clean
make html