This is a mixed notebook and data repository for retrieving cancer data for Project Cognoma. Currently, all data is from the TCGA Pan-Cancer collection of the UCSC Xena Browser.
The data acquisition and analysis is executing by running Jupyter notebooks in the following order:
1.TCGA-download.ipynb— download and compress TCGA datasets.2.TCGA-process.ipynb— convert downloaded TCGA datasets into sample × gene matrixes.
The execute.sh script executes the notebooks in order.
After installing and activating the environment, run with the command bash execute.sh from the repository's root directory.
The repository contains the following directories:
download— contains files retrieved from an external location whose content is unmodified. Large downloaded files are tracked using Git LFS. Associated metadata files are also retained for versioning.data— contains generated datasets. The complete matrix files are not currently tracked due to file size, but randomly-subsetted versions are available for development use (seedata/subset).
The complete datasets created by this repository (data/expression-matrix.tsv.bz2 and data/mutation-matrix.tsv.bz2) are uploaded to figshare.
Since this is a manual process, check the figshare REFERENCES section to see which commit these datasets derive from.
In other words, the latest version on figshare may lag behind this repository.
This repository uses conda to manage its environment, which is named cognoma-cancer-data.
The required packages and versions are listed in environment.yml.
If as a developer, you require an additional package, add it to environment.yml.
The following commands install and activate the environment:
# Create or overwrite the cognoma-cancer-data conda environment
conda env create --file=environment.yml
# Activate the conda environment (assumes conda >= 4.4)
conda activate cognoma-cancer-dataThis repository is dual licensed as BSD 3-Clause and CC0 1.0, meaning any repository content can be used under either license. This licensing arrangement ensures source code is available under an OSI-approved License, while non-code content — such as figures, data, and documentation — is maximally reusable under a public domain dedication.