ADToolbox

From raw amplicon reads to a calibrated anaerobic digestion model.

ADToolbox is developed in the Chan Lab at Colorado State University. It connects metagenomics evidence, curated reaction and feed databases, and dynamic ADM1 / e-ADM simulations into one reproducible Python and command-line workflow.

Anaerobic digestion models such as ADM1 lump the microbial community into a handful of guilds whose initial biomass you are expected to guess. ADToolbox replaces that guess with measurement: it takes 16S amplicon data, maps it through GTDB and a curated enzyme-to-reaction database, and produces the microbial COD allocation the model needs.

📖 Read the documentation

Install

pip install adtoolbox               # base install, Python 3.11+
pip install "adtoolbox[optimize]"   # + parameter-tuning backends
pip install "adtoolbox[dashboard]"  # + interactive Dash/Escher visualization

Or use the container, which bundles fastp, VSEARCH, MMseqs2, and the SRA Toolkit:

docker run --rm parsaghadermazi/adtoolbox:latest adtoolbox --help

See the installation guide for source installs, extras, external tools, and HPC notes.

Quick start

# 1. Download the reference databases
adtoolbox database download-all-databases --output-dir ./database

# 2. Turn amplicon samples into microbial COD allocations
adtoolbox metagenomics process \
  --input ./samples.tsv --input-type sra --assay amplicon \
  --output-dir ./process --sra-dir ./sra \
  --database-dir ./database --execute

# Use the direct functional route for shotgun reads
adtoolbox metagenomics process \
  --input ./shotgun_samples.tsv --input-type reads --assay shotgun \
  --output-dir ./process --database-dir ./database --execute

# 3. Simulate
adtoolbox adm e-adm --models-json reference_data/models.json --report csv

The full walkthrough is in the Quickstart.

Documentation map

Page	Contents
Quickstart	Install, databases, first simulation, first pipeline run.
Metagenomics Pipeline	Sample tables, execution profiles, Slurm, output schemas.
ADM Models	Input contracts, stoichiometry, rate laws, inhibition terms.
Parameter Tuning	Fitting kinetic parameters to experimental data.
CLI	Every command and option.
Python API	Generated reference for all modules.

Try it without installing

The example notebooks run on Binder. Note that Escher map visualization is not available there.

Building the docs locally

pip install -r docs/requirements.txt
mkdocs serve

Contact

Developed in the Chan Lab at Colorado State University.

Parsa Ghadermazi	[email protected]
Ethan Rimelman	[email protected]
Siu Hung Joshua Chan (PI)	[email protected]