1. Clone repository
2. Install dependecies: run conda env create -f environment.yml
3. Start conda environment: run source activate <env>
4. Install jellyfish You do not need to install the python binding for jellyfish, just the command line tool
5. Install kmerprediction: python setup.py install
6. Verify that everything is working: run nose2 this will run all of the tests, they should all pass.

• File/filepath does not exist
  • Update the filepaths in kmerprediction/constants.py to point to the correct locations on your machine
  • You will have to rerun python setup.py install in order for the changes to take effect.
• Import error stating that python can't find the module lmdb
  • run pip install lmdb with the conda environment activated.
• Error like:

  File "/home/user/miniconda3/envs/kmer/lib/python2.7/site-packages/hyperopt/pyll/base.py", line 715, in toposort
     assert order[-1] == expr
  TypeError: 'generator' object has no attribute '__getitem__'
  

  • Downgrade networkx from version 2.0 to version 1.1.

Update kmerprediction/constants.py so that OMNILOG_FASTA, OMNILOG_DATA,GENOME_REGION_TABLE, and ECOLI point to the correct locations on your machine. You will have to rerun python setup.py install in order for the changes to take effect.

• OMNILOG_FASTA: Path to directory containing the Omnilog fasta files,
• OMNILOG_DATA: Path to the wide_format_header.txt file containing the Omnilog AUC data.
• GENOME_REGION_TABLE: Path to the binary table containing genome region presence/absence data for the US/UK e. Coli samples.
• ECOLI: Path to the directory containing the US/UK e. Coli fasta files.

Run snakemake -s validation.smk

Run snakemake -s omnilog.smk

See kmerprediction/README.md