A Python implementation of the NPC toolchain for the import, quality-control, and preprocessing of metabolic profiling datasets.
Imports:
- Peak-picked LC-MS data (XCMS, Progenesis QI, & Metaboscape)
- Raw NMR spectra (Bruker format)
- Targeted datasets (TargetLynx, Bruker BI-LISA & BI-Quant-Ur)
Provides:
- Batch & drift correction for LC-MS datasets
- Feature filtering by RSD and linearity of response
- Calculation of spectral line-width in NMR
- PCA of datasets
- Visualisation of datasets
Exports:
- Basic tabular csv
- ISA-TAB
Tutorials:
- Available at nPYc-toolbox-tutorials, see below.
For full installation instructions see Installing the nPYc-Toolbox
To install via pip, run:
pip install nPYc
To install from a local copy of the source, simply navigate to the main package folder and run:
python setup.py install
Alternatively, using pip and a local copy of the source:
pip install /nPYC-toolboxDirectory/
To update the current installed version use:
pip install --upgrade nPYc
Installation with pip allows the usage of the uninstall command
pip uninstall nPYc
Documentation is hosted on Read the Docs.
Documentation is generated via Sphinx Autodoc, documentation markup is in reStructuredText.
To build the documentation locally, cd into the docs directory and run:
make html
To clear the current documentation in order to rebuild after making changes, run:
make clean
A repository containing exemplar datasets and Jupyter notebook tutorials to demonstrate the application of the nPYc-Toolbox for the preprocessing and quality control of LC-MS, NMR and targeted NMR (Bruker IVDr) metabolic profiling data is available for download from nPYc-toolbox-tutorials.
For new users, we strongly recommend downloading these tutorials, which provide detailed worked examples with links to relevant documentation.
Source management is git-flow-like - no development in the master branch! When making a change, create a fork based on develop, and issue a pull request when ready.
When merging into the develop branch, all new code must include unit-tests, all tests should pass, and overall code-coverage for the toolbox should not drop.
When merging from develop (or hotfix branches) into release, ensure:
- All references to the debugger are removed
- All paths are relative and platform agnostic
- All tests pass
Unit testing is managed via the unittest framework. Test coverage can be found on codecov.io.
To run all tests, cd into the Tests directory and run:
python -m unittest discover -v
Individual test modules can be run with:
python -m `test_filename` -v