Skip to content

Example Notebooks

The Examples/ directory in the repository holds executable notebooks for each major workflow. Two of them are also rendered here as static walkthroughs.

Available notebooks

  • modeling.ipynb

    Build a new ADM model or load an existing one, then solve, inspect, and visualize it. The best starting point if you care about simulation rather than sequencing.

    Rendered walkthrough · Source

  • parameter_tuning.ipynb

    Reproduces the figures from the ADToolbox paper, including predicting AD products from 16S data at time zero, and walks through the full parameter-tuning workflow.

    Rendered walkthrough · Source

  • metagenomics_notebook.ipynb

    The sequence-analysis path in detail, covering both 16S amplicon and shotgun data and the computational steps that connect them to the model.

    Source

  • pipeline.ipynb

    A compact driver for the batch metagenomics process pipeline from Python, useful as a template for your own runs.

    Source

Running them

git clone https://github.com/chan-csu/ADToolbox.git
cd ADToolbox
pip install -e ".[optimize,dashboard]"
pip install jupyterlab
jupyter lab Examples/

Binder

No installation required. Escher map visualization is not available on Binder.

Supporting data

The notebooks read from files that ship with the repository:

Path Contents
Examples/ADM_Parameters/ Individual e-ADM parameter JSON files.
Examples/Studies/ Metagenomics study metadata and experimental data references.
Examples/feed_db.tsv A small feed database.
Examples/toy_model.json, Examples/Toymap.json A minimal model and its Escher map.
reference_data/models.json ADM1 and e-ADM parameter sets keyed by model name.

Notebooks that need databases

The metagenomics notebooks expect the reference databases. Run adtoolbox database download-all-databases --output-dir ./database first, and point the notebook's database_dir at that directory. See the Quickstart.