PyDrugLogics

Overview

PyDrugLogics is a Python package designed for constructing, optimizing Boolean Models and performs in-silico perturbations of the models.

Core Features

• Construct Boolean model from .sif file
• Load Boolean model from .bnet file
• Optimize Boolean model
• Generate perturbed models
• Evaluate drug synergies

Installation

PyDrugLogics can be installed via PyPi, Conda, or directly from the source.

Install PyDrugLogics from PyPI

The process involves two steps to install the PyDrugLogics core package and its necessary external dependencies.

1. Install PyDrugLogics via pip

pip install pydruglogics

2. Install External Dependency

pip install -r https://raw.githubusercontent.com/druglogics/pydruglogics/main/requirements.txt

This will install the PyDrugLogics package and handle all dependencies automatically.

Install PyDrugLogics via conda

Note: CoLoMoTo conda integration is ongoing.

conda install szlaura::pydruglogics

Install from Source

For the latest development version, you can clone the repository and install directly from the source:

git clone https://github.com/druglogics/pydruglogics.git
cd pydruglogics
pip install .
pip install -r requirements.txt

CoLoMoTo Notebook envionment

PyDrugLogics is available in the CoLoMoTo Docker and Notebook starting from version 2025-01-01.

Setup CoLoMoTo Docker and Notebook

1. Install the helper script in a terminal:

    pip install -U colomoto-docker

2. Start the CoLoMoTo Notebook (a specific tag can also be given):

    colomoto-docker    # or colomoto-docker -V 2025-01-01

3. Open the Jupiter Notebook and navigate to the tutorials folder to find the PyDrugLogics folder hosting the pydruglogics tutorial notebook.

See more about the CoLoMoTo Docker and Notebook in the documentation.

Documentation

For full PyFrugLogics documentation, visit the GitHub Documentation.

Quick Start Guide

Here's a simple example to get you started:

from pydruglogics.model.BooleanModel import BooleanModel
from pydruglogics.input.TrainingData import TrainingData
from pydruglogics.input.Perturbations import Perturbation
from pydruglogics.input.ModelOutputs import ModelOutputs
from pydruglogics.execution.Executor import execute

# Initialize train and predict
model_outputs_dict = {
        "RSK_f": 1.0,
        "MYC": 1.0,
        "TCF7_f": 1.0
    }
model_outputs = ModelOutputs(input_dictionary=model_outputs_dict)

observations = [(["CASP3:0", "CASP8:0","CASP9:0","FOXO_f:0","RSK_f:1","CCND1:1"], 1.0)]
training_data = TrainingData(observations=observations)


drug_data = [['PI', 'PIK3CA', 'inhibits'],
            ['PD', 'MEK_f', 'activates'],
            ['CT','GSK3_f']]
perturbations = Perturbation(drug_data=drug_data)


boolean_model = BooleanModel(file='./ags_cascade_1.0/network.bnet', model_name='test', mutation_type='topology',
                                  attractor_tool='mpbn', attractor_type='trapspaces')

observed_synergy_scores = ["PI-PD", "PI-5Z", "PD-AK", "AK-5Z"]


ga_args = {
        'num_generations': 20,
        'num_parents_mating': 3,
        'mutation_num_genes': 3,
        'fitness_batch_size': 20
}

ev_args = {
        'num_best_solutions': 3,
        'num_of_runs': 30,
        'num_of_cores': 4
}


train_params = {
        'boolean_model': boolean_model,
        'model_outputs': model_outputs,
        'training_data': training_data,
        'ga_args': ga_args,
        'ev_args': ev_args
}

predict_params = {
        'perturbations': perturbations,
        'model_outputs': model_outputs,
        'observed_synergy_scores': observed_synergy_scores,
        'synergy_method': 'bliss'
}

# run train and predict
execute(train_params=train_params, predict_params=predict_params)

For a more detailed tutorial, please visit the documentation or the tutorial.

Citing PyDrugLogics

If you use PyDrugLogics, please cite the paper:

Flobak, Å., Zobolas, J. et al. (2023): Fine tuning a logical model of cancer cells to predict drug synergies: combining manual curation and automated parameterization. DOI: 10.3389/fsysb.2023.1252961

@Article{druglogics2023,
  title = {Fine tuning a logical model of cancer cells to predict drug synergies: combining manual curation and automated parameterization},
  author = {Flobak, Å., Zobolas, J. and Other Authors},
  journal = {Frontiers},
  year = {2023},
  month = {nov},
  doi = {10.3389/fsysb.2023.1252961},
  url = {https://www.frontiersin.org/journals/systems-biology/articles/10.3389/fsysb.2023.1252961/full},
}