This repo hosts the code for the quantitative image analysis and Ordinary Differential Equation (ODE) modelling in the paper Desensitisation of Notch signalling through dynamic adaptation in the nucleus by Viswanathan, Hartmann and colleagues. All code was written by Jonas Hartmann.

• Notebooks/ is where most of the important code is.
  • DEV: Initial development of image quantification pipelines.
  • RUN: Batch execution of the image quantification pipelines.
  • ANA: Downstream analysis of quantified image data.
  • MOD: The ODE modelling and its analysis.
• optonotch/ is a package into which some frequently used functions have been refactored.
  • optonotch.utilities: Various utilities used in various places.
  • optonotch.modeling: Data loading, input functions and loss functions for the modelling.
• Data/ houses the data used.
  • Images/: Input image data. Required only for DEV and RUN notebooks. Not hosted on GitHub due to file size; please contact Jonas Hartmann for more information.
  • Measurements/: Derived values. Required for ANA and MOD notebooks. Included in this repository.
• Figures/: Figures generated by the code.
  • These are "raw" plots used for figure construction; composition, axis labels and other annotations may be slightly different in the published paper figures.
• Illustrations/ contain some images used in markdown annotations in some places.

1. Raw images were prepared by z-projection where appropriate.

2. RUN notebooks ingest images and produce measurements.

3. ANA notebooks produce figures from these measurements.

• ANA - sim spot detection.ipynb also generates and saves a cleaned version of some measurements for use in modelling!

• ANA - nuclear nicd bleaching.ipynb also infers the bleaching constant of mCherry for use in fitting the NICD import-export models.

• ANA - nuclear nicd measurement.ipynb also fits NICD import-export models for use in modelling!

4. MOD notebooks fit the various models that were tested and produce figures from them.

• Python 3.7.3 (I used and recommend the Anaconda distribution)
• jupyter==7.6.1 and ipython==7.6.1
• Scientific python stack, most importantly:
  • jupyter==1.0.0
  • numpy==1.16.4
  • scipy==1.2.1
  • pandas==0.24.2
  • matplotlib==3.1.0
  • scikit-image==0.15.0
  • sympy==1.4
• Oh, and dill==0.3.3

• The study's lead author is Stefano De Renzis at . Please contact him to request any non-computational materials, resources or reagents,
• For questions regarding the code and computational resources, please contact Jonas Hartmann at or open an issue on GitHub. Note that I cannot promise support for any use cases other than direct reproduction of the study's results.