Data preparation

The multimodal experiment combines two data sources from the same cohort. Both have already been documented in the single-modality tutorials. Follow the steps below to ensure both are ready.

Microbiome profiles

Download the raw metadata and taxonomic profiles, then build the hierarchical TSV:

1. Run the download script as described in Data preparation (microbiome tutorial).
2. Run prepare_hierarchical_profiles.py as described in Preparing hierarchical profiles.

The output file is example/IBD/data/FRANZOSA_IBD_2019_profiles_hierarchical.tsv.

Metabolomics data

Download the raw metabolomics matrix and rename the compound columns:

1. Run the download script as described in Data preparation (metabolomics tutorial).
2. Run prepare_mtb_data.py as described in Preparing the metabolomics file.

The output file is example/IBD/data/metabolomics_data.tsv.

Metadata

Both modalities share the same metadata file (example/IBD/data/metadata.tsv) downloaded in the first step. The Sample column is the shared key that aligns samples across modalities.