Offline Usage

FLUTE requires Python installed on local machine (macOS, Linux and Windows are all supported). If users want to explore the interactive notebook we provided locally, Jupyter notebook installation is also required.

  1. Clone the FLUTE repository, to your computer.

git clone https://github.com/pitt-miskov-zivanov-lab/FLUTE.git

  1. Navigate into the directory, install FLUTE and its python dependencies.

cd FLUTE
pip install -e .

  1. Use either in terminal or via notebook.

3.1 In terminal: change working directory to src/ folder by cd src/, first unzip a large file unzip ../supplementary/oa_file_list.txt.zip -d "../examples/input/", and then see what parameters are needed in run_FLUTE.py via helper

python run_FLUTE.py -h
>> usage: run_FLUTE.py [-h] [-g] [-p] [-q [PROT_QUERY]] [-r] username password host db_name input_file output_path

That’s being said, six ordered positional arguments are required: username password host db_name input_file output_path, some optional flag/arguments can be appended to support additional features:

More Arguments

Examples

Descriptions

No optional ones

python run_FLUTE.py root 12345678 localhost flute ../examples/input/example.xlsx ../examples/output/example

Filter and score input interactions

With -g or --drop_duplicates

python run_FLUTE.py root 12345678 localhost flute ../examples/input/example.xlsx ../examples/output/example -g

Also output duplicate interactions in input file

With -p or --keep_recent_pps

python run_FLUTE.py root 12345678 localhost flute ../examples/input/example.xlsx ../examples/output/example -p

Also output interactions published in papers of recent years

With -q [PROT_QUERY] or --prot_query [PROT_QUERY]

python run_FLUTE.py root 12345678 localhost flute ../examples/input/example.xlsx ../examples/output/example -q P00533,P03386

Also find interactions related to these given protein IDs

With -r or --add_related_ints

python run_FLUTE.py root 12345678 localhost flute ../examples/input/example.xlsx ../examples/output/example -r

Also find additional interactions related that are in the same paper set

Combine -g -p or any option

python run_FLUTE.py root 12345678 localhost flute ../examples/input/example.xlsx ../examples/output/example -g -p

Do both -g and -p, or any provided

3.2 Via notebook: Run the provided notebook after cd examples/ from FLUTE root directory. Detailed instructions can be found in markdown cells inside the notebook or at this page

jupyter notebook examples/use_FLUTE.ipynb

  1. Prepare input files and interpret output files

4.1 Input and parameters include:

  • a .xlsx file containing the machine reading interactions, best in BioRECIPES interaction format, see example. At the minimum requirements, it shall have the following headers:

Regulated Name

Regulated ID

Regulated Type

Regulator Name

Regulator ID

Regulator Type

Paper IDs

  • a score tuple (es, ts, ds), three positive numbers, denoting thresholds of escore, tscore, dscore. Default value is set to be (0,0,0), update it in main() (run in terminal) or in Cell 18 and 27 (run with notebook).

  • If using the feature of returning recent papers, integer x is needed to specify how many years to return interactions published in recent papers. Default is 5, update it in main() (run in terminal) or in Cell 38 (run with notebook).

  • If using the feature of querying protein IDs and returning related interactions, string of protein IDs is needed, e.g., P00533,P03386, supported after -q (run in terminal) or in Cell 32 (run with notebook).

4.2 Outputs (including optional ones) include:

  • list of reading interactions that pass filtration, see example

  • the filtration scores for those filtered interactions, see example

  • OPTIONAL: duplicated reading interactions and their occurrences, see example

  • OPTIONAL: interactions published in recent x years, see example

  • OPTIONAL: interactions related to query protein IDs, see example

  • OPTIONAL: interactions that are in the same papers as the input file, see example