Skip to content

MathOnco/PARPi_Model

Repository files navigation

Strobl et al (2024). To modulate or to skip: De-escalating PARP inhibitor maintenance therapy in ovarian cancer using adaptive therapy

This repository contains the code and data for our publication Strobl et al (2024). To modulate or to skip: De-escalating PARP inhibitor maintenance therapy in ovarian cancer using adaptive therapy, Cell Systems xxx, available here [xxx]. A pre-print of our manuscript is available on the bioRxiv [2].

Visual summary of our study. Using an integrative and iterative process we developed and validated a mathematical model of the PARPi response in ovarian cancer. Leveraging this model we explored different adaptive therapy algorithm, finding that modulations based strategies are superior to skipping. This is due to a delay in the cell kill and a diminishing dose response. Preliminary in vivo experiments confirm that adaptive modulation can control the tumor as well as continuous therapy whilst halving drug use

Requirements

A full list of the Python packages used in this project can be found in requirements.txt. To recreate the virtual environment, run:

$ conda create --name <envname> --file requirements.txt
$ source <env_name>/bin/activate

For further details, see here

Data

Both the raw and processed data files can be found in the data folder. These contain all the confluence vs time data used to calibrate and validate the models shown in the paper (in vitro and in vivo). The data processing steps are documented in jnb_dataProcessing.ipynb:

  • continuousTreatmentDf_raw.csv contains the data for the experiments in which we treated cells continuously at different doses and from different starting densities.
  • continuousTreatmentDf_cleaned.csv contains the cleaned continuous treatment data that we used for model fitting/testing (see jnb_dataProcessing.ipynb for details of post-processing).
  • intermittentTreatmentDf_oc3_raw.csv and intermittentTreatmentDf_oc3_raw.csv contains the raw data for the experiments in which we treated cells for some time and then withdrew treatment.
  • intermittentTreatmentDf_cleaned.csv contains the cleaned intermittent treatment data that we used for model fitting/testing (see jnb_dataProcessing.ipynb for details of post-processing).
  • mouseDataDf_oc3.csv contains the volume vs time data from the in vivo experiment.
  • sweep_mod_vs_skipping.csv contains the simulation data from comparing modulation-based and skipping based strategies in Figure 7a.

Analysis

For each results figure in the manuscript we have created a separate jupyter notebook which houses the code to re-create this figure. These are named jnb_figure2.ipynb etc. and contain further explanations within.

In case of questions or comments, feel free to reach out to me at anytime.

References

  • [1] xxx
  • [2] Strobl, M. et al. (2023). Adaptive therapy for ovarian cancer: An integrated approach to PARP inhibitor scheduling. bioRxiv 2023.03.22.533721, doi:10.1101/2023.03.22.533721.