FEV-KEGG 1.1.4

FEV@KEGG

FEV@KEGG allows for easy analysis of metabolic networks of organisms in KEGG (Kyoto Encyclopedia of Genes and Genomes). Read the API documentation here: https://fev-kegg.readthedocs.io

Install

Use pip to install FEV@KEGG and to automagically install all dependencies: pip install FEV_KEGG

If you are on Python 3.4, you will have to use pip install FEV_KEGG[python34] to pull in the backported typing package.

Where to start?

After successful installation, you might want to take a look at the “experiment” scripts in FEV_KEGG/Experiments. These scripts consecutively involve more and more functionality of this library. They were used during development, step by step adding and testing another layer of functionality or abstraction. Therefore, they might be useful to you in learning how to use this very functionality.

Also, take a look at the API documentation: https://fev-kegg.readthedocs.io

If any questions remain, feel free to report an issue: https://github.com/ryhaberecht/FEV-KEGG/issues

Dependencies

These are automatically installed by pip.

• Python 3.4+

• NetworkX

• anytree

• jsonpickle

• tqdm

• BeautifulSoup

• retrying

• appdirs

• typing (for Python 3.4 only)

Optional Dependencies

If you want to draw a graph to an image file:

• PyGraphviz

• Graphviz (non-python software you will have to install manually!)

Use pip install FEV_KEGG[draw_image].

If you want to draw a graph in a pop-up window:

• Matplotlib

Use pip install FEV_KEGG[draw_window]

Exporting to GraphML or GML works without any optional dependencies.

Included Dependencies

These have been partially copied into this project to avoid unnecessarily big dependencies and allow for minor changes.

• Bio.KEGG from Biopython in lib.Biopython.KEGG

Recommendations

• SSD

When handling 500 organisms from KEGG at once:

• 64 bit operating system

• 4 GB RAM

• 20 GB disk space for cache

When handling all ~5000 organisms in KEGG at once:

• 64 bit operating system

• 12 GB RAM

• 100 GB disk space for cache

Developer’s System

• cPython 3.4.6

• x86-64 Linux (OpenSUSE Leap 42.3)

• 16 GB RAM

• 1 CPU, 2 Cores, 4 Threads

• SSD

Caching

• The cache directory path is set up in the ‘settings.py’ file on the top level of the project. Per default, it points to your user’s cache directory as defined by your OS.

  • Linux/Unix: ~/.cache/FEV-KEGG

  • OS X: ~/Library/Caches/FEV-KEGG

  • Windows: C:\Users\username\AppData\Local\ryh\FEV-KEGG\Cache

• All downloads from KEGG are cached automatically. Also, basic graphs are cached by organism. These default cachings alone can grow the cache directory to 100 GB size!

• You can cache any function’s result using the @cache decorator, see FEV_KEGG.KEGG.File.cache. Watch out to remember the path and file name and not to overwrite any other cached files.

• To cause a download of the newest version of data from KEGG, you have to delete the cached file manually. Have a look inside the ‘cache’ folder, file paths and names should be self-explanatory.

• On Linux with supporting file systems, disabling atime (file access time) for the cache directory and all its contents might improve performance: sudo chattr -R +A ~/.cache/FEV-KEGG

1.0.0 (2018-08-09)

• Initial release.

1.1.1 (2018-08-18)

• Bug-fix: When choosing the type of redundancy in the RedundancyType enum, TARGET_FLEXIBILITY and SOURCE_FELXIBILITY used to always equal to FLEXIBILITY, due to the scriptiness of Python. Now they are treated differently.

• Added partial redundancy to the the RedundancyType enum, plus logic for the Redundancy and RedundancyContribution classes.

• Added export of dictionaries containing Neofunctionalisation/FunctionChange/Element objects into a formatted HTML file.

• Added three more experiments involving the above.

1.1.2 (2018-08-18)

• Bug-fix: Added missing majority percentage parameter in several experiments.

• Added new experiment.

• Improved sorting of neofunctionalisations. Now they sort by EC number first.

• Improved sorting for enzymes. Now they sort by EC number first.

• Improved sorting for EC numbers. Now they sort naturally, not lexicographically.

1.1.3 (2019-01-07)

• Bug-fix: Improved parsing and error handling of gene description files.

• Bug-fix: Fixed deprecated API-calls in older experiments.

• Bug-fix: Export of GraphML failed due to None values.

• Added experiment 45.

• Fixed typos.

• Master’s thesis release!

1.1.4 (2022-05-20)

• Bug-fix: non-existing attribute

• Bug-fix: KEGG changed attributes of gene type