Multi-Omics Predictive Modelling of Cervicovaginal Microenvironment

This repository accompanies the preprint: "Integration of multi-omics data improves prediction of cervicovaginal microenvironment in cervical cancer" (citation below). It includes all published analyses, namely the data analysis and modelling in notebook a-modelling-HPV.ipynb and the analysis of the interaction of microbiome and metabolome in the notebook b-mmvec-HPV.ipynb.

To reproduce the published results please follow the below setup instructions (they are unique for each of the two notebooks).

Setup for a-modelling-HPV.ipynb

• To run this notebook you should setup a conda environment with the qiime2-2021.4 distribution installed within:

wget https://data.qiime2.org/distro/core/qiime2-2021.4-py38-osx-conda.yml
conda env create -n hpv_modelling --file qiime2-2021.4-py38-osx-conda.yml
rm qiime2-2021.4-py38-osx-conda.yml
conda activate hpv_modelling

• Within the conda environment install all required packages with (usage of python -m pip to ensure that the package is pip installed in the conda environment and not elsewhere):

python -m pip install git+https://github.com/bokulich-lab/RESCRIPt.git
conda install -c conda-forge -c r --file requirements-modelling.txt

• Have fun recreating our published results :).

Setup for b-mmvec-HPV.ipynb

• To run this notebook you should setup a conda environment with an older version of the QIIME2 distribution installed within, namely 2020.6. This is required as the used plugin mmvec is currently only supported until this version.

wget https://data.qiime2.org/distro/core/qiime2-2020.6-py36-osx-conda.yml
conda env create -n hpv_mmvec --file qiime2-2020.6-py36-osx-conda.yml
rm qiime2-2020.6-py36-osx-conda.yml
conda activate hpv_mmvec

• Within the activated conda environment install the required dependency mmvec as:

python -m pip install install git+https://github.com/biocore/mmvec.git
qiime dev refresh-cache

• (Optional:) If you want to have a nice jupyter notebook experience feel free to also install the following dependencies:

conda install -c conda-forge --file requirements-mmvec.txt

• Happy reproduction of the results! :)

Contact

In case of questions or comments feel free to raise an issue in this repository.

Citation

If you use code, data, or ideas from this repository, please cite:

Bokulich NA, Łaniewski P, Adamov A, Chase DM, Caporaso JG, Herbst-Kralovetz MM (2022) Multi-omics data integration reveals metabolome as the top predictor of the cervicovaginal microenvironment. PLoS Comput Biol 18(2): e1009876. https://doi.org/10.1371/journal.pcbi.1009876