NOTE !!!

This has been temporarily moved to: https://github.com/Kizielins/q2-predict-dysbiosis/tree/master

QIIME 2 plugin for calculating dysbiosis score from gut microbiome data. A greater score indicates better health.

Installation

To install the most up to date version of the plugin:

• Install and activate conda environment with QIIME 2 (see docs), e.g. for Linux 64-bit. Note that this plugin is currently only available in the QIIME 2 dev version. This should only take a few minutes:

  wget https://raw.githubusercontent.com/qiime2/environment-files/master/latest/staging/qiime2-latest-py38-linux-conda.yml
  conda env create -n qiime2-dev --file qiime2-latest-py38-linux-conda.yml
  rm qiime2-latest-py38-linux-conda.yml
  conda activate qiime2-dev
  qiime info
  

Note that the plugin was tested with qiime2-2022.2 .

• Fetch the repository and go to main folder:

  git clone https://github.com/Kizielins/q2-predict-dysbiosis.git
  cd q2-predict-dysbiosis
  

• Install plugin:

  pip install -e .
  python setup.py install
  

• Test plugin e.g.: qiime predict-dysbiosis --help

Input prep:

Sample inputs can be found in the "test_data" folder.

• Taxonomy table: standard QIIME 2 *qza feature table, collapsed to species level, with removed "s__" and underscores instead of spaces (ie "Escherichia_coli")
• Stratified pathways table: standard QIIME 2 *qza feature table, produced by HUMAnNN, collapsed to species level, with underscores instead of spaces (ie ANAEROFRUCAT-PWY:_homolactic_fermentation|g__Citrobacter.s__Citrobacter_freundii)
• Unstratified pathways table: standard QIIME 2 *qza feature table, produced by HUMAnNN, with underscores instead of spaces (ie AEROBACTINSYN-PWY:_aerobactin_biosynthesis)
• Metadata: standard QIIME 2 metadata format, with "id" and columns representing sample IDs and labelling.

The values in all tables should be expressed as relative abundance.

Predict dysbiosis index

Usage: qiime predict-dysbiosis calculate-index [OPTIONS]
Dysbiosis index predicts the gut microbiome health index for each sample in the abundance table.

Inputs:

--i-table ARTIFACT FeatureTable[RelativeFrequency]
Abundance table artifact with taxonomy collapsed to species level.

--i-pathways-stratified ARTIFACT FeatureTable[RelativeFrequency]
Abundance table artifact with stratified pathways.

--i-pathways-unstratified ARTIFACT FeatureTable[RelativeFrequency]
Abundance table artifact with unstratified pathways.

Outputs:

--o-dysbiosis-predictions ARTIFACT SampleData[AlphaDiversity]
Predicted dysbiosis index in tabular form.

Predict and visualize dysbiosis index

Usage: qiime predict-dysbiosis calculate-index-viz [OPTIONS]
Dysbiosis index predicts the gut microbiome health index for each sample in the abundance table.

Inputs:

--i-table ARTIFACT FeatureTable[RelativeFrequency]
Abundance table artifact with taxonomy collapsed to species level.

--i-pathways-stratified ARTIFACT FeatureTable[RelativeFrequency]
Abundance table artifact with stratified pathways.

--i-pathways-unstratified ARTIFACT FeatureTable[RelativeFrequency]
Abundance table artifact with unstratified pathways.

--m-metadata-file ARTIFACT
Metadata file.

Outputs:

--o-index_results ARTIFACT SampleData[AlphaDiversity]
Predicted dysbiosis index in tabular form.

--o-index_results ARTIFACT Visualization
Predicted dysbiosis index visualization file.

Sample use

Calculate index:

qiime predict-dysbiosis calculate-index --i-table test_files/taxonomy.qza --i-pathways-stratified test_files/pathways_stratified.qza --i-pathways-unstratified test_files/pathways_unstratified.qza --o-dysbiosis-predictions results.qza

Calculate and visualize index:

qiime predict-dysbiosis calculate-index-viz --i-table test_files/taxonomy.qza --i-pathways-stratified test_files/pathways_stratified.qza --i-pathways-unstratified test_files/pathways_unstratified.qza --m-metadata-file test_files/metadata.txt --o-index-results results.qza --o-index-plot visualization.qzv

Both commands should take a few minutes to run, depending on the size of your input (~30s per sample).

Original publication / citation

If you want to learn more about this method, or to cite it, please refer to our article: https://www.biorxiv.org/content/10.1101/2023.12.04.569909v4

A full script to reproduce all figures in the article will be available shortly.

Acknowledgements

We would like to acknowledge the Authors of the q2-health-index plugin, whose scripts formed the foundation of our work.