figure2

figure2/

This directory contains the data and code to generate Figure 2 in the paper (generating executable models from KEGG PATHWAYS).

About the files

biomass_models/

This directory contains the biomass models used in the paper.

KEGG_erbb_MCF7 & KEGG_erbb_MDAMB231

Includes the model files for the MCF7 and MDAMB231 models, respectively. Both models were generated using the same Text2Model file (KEGG_erbb_text2model.txt), but fitted to different experimental data.

KEGG2Model/

This directory contains the code necessary to convert KEGG PATHWAYS to Text2Model files, as well as conducting co-occurrence analysis on the PubTator data.

Contents

• convert_network.py: Converting KEGG PATHWAYS to Text2Model files.
• cooccurrence_analysis: Conducting co-occurrence analysis on the PubTator data.
• KGML_parser.py: Parsing KGML files.
• weight_visualizer.py: Visualizing weights of networks.

pubtator/

This directory contains the PubTator data used in the paper.

Although the script to reproduce the results can be found as prepare_data.py and process_data.py, this process can be data-intensive and time-consuming. Furthermore, the PubTator data that is used in the paper is the Jan. 2022 release, which seems to be no longer available. Therefore, the processed data is included as a compressed json file (pubtator_data.json.gz).

Reproducing the results

Network Visualizations

The network visualizations can be reproduced using the KEGG_script.py script. The script will create and output the results under out/.

Simulation Results

The experimental data used to estimate the model parameters were obtained from a previous study (Imoto et al., 2020). The data can be found in the observable.py file in both model directories:

class Observable(DifferentialEquation):
    ...
    def set_data(self) -> None:
        self.experiments[self.obs_names.index("Phosphorylated_AKT")] = {
            "EGF": [0.0, 0.242, 0.087, 0.088, 0.082, 0.045, 0.017, 0.043],
            "HRG": [0.0, 0.976, 1.0, 0.96, 0.876, 0.836, 0.77, 0.719],
        }
        ...

    def get_timepoint(self, obs_name: str) -> List[int]:
        if obs_name in self.obs_names:
            return [0, 5, 15, 30, 45, 60, 90, 120]

The code used to estimate the parameters and generate the plots can be found in the /biomass_models/biomass_scipt.py file as an executable script. Reproducing the results can take up to a couple of days depending on the available computational resources.