MGP prediction
==============

.. automodule:: MGP prediction 
    :members: 
    :undoc-members: 
    :show-inheritance: 
        
**How to install**

------

It works on Linux operating system and has been tested on Ubuntu 16.04. The runtimes below are generated using a computer with the specs (32 GB RAM, Intel(R) Core(TM) i7-4790 CPU @ 3.60 GHz) and internet of speed around 80 Mbit/s.

1. Clone the repository

    .. code-block:: bash

        git clone https://bitbucket.org/kaistsystemsbiology/mgp_prediction.git
        

2. Create and activate a conda environment


    .. code-block:: bash

        cd mgp_prediction
        conda env create -f environment.yaml    
        source activate MGPprediction

which should install in about 2 minutes

------

**Implementation**
1. Create .csv file for flux (e.g.,exampleFlux.csv) and mutation (e.g.,exampleMutation.csv). Please referred to the format of each file in exampleInput directory.

2. Run MGP prediction. The following arguments are required: -o/--output_dir, -f/--flux, -mut/--mutation

    **Example**
    .. code-block:: bash
        mkdir output
        python runMGPprediction.py -o ./output -f ./exampleInput/exampleFlux.csv -mut ./exampleInput/exampleMutation.csv
    The output file "predictedMGPs.csv" will be generated in output directory after computation.
    The run time for this source code on example input is about 90 seconds.

3. The output file is in .csv format and consists of four columns (i.e., Target gene, Target metabolite and its MetaNetX ID, and Target pathway) which represent information of predicted MGPs.

------

**Reference**
Lee, G., Lee, S.M., Lee, S. et al. Prediction of metabolites associated with somatic mutations in cancers by using genome-scale metabolic models and mutation data. Genome Biol 25, 66 (2024). https://doi.org/10.1186/s13059-024-03208-8

-------