Keratins coordinate tissue spreading by balancing spreading forces with tissue material properties data repository.
This repository contains the data analysis notebooks and scripts used to generate figures for the Naik et al. 2026 manuscript. It is aimed to organize to keep raw/derived data in addition to the notebooks that plot the figure panels and reusable scripts, so figure panels used in the manuscript can be reproduced consistently.
This repo is mainly maintained by me, Dr. Suyash Naik. In case of any questions please reach out to me by email provided below. The vertex model accompanying this paper and its associated data was generated by Dr. Yann-Edwin Keta and Dr. Silke Henkes. Their main repo is hosted on Yann's git here.
For tissues to spread, they must deform while staying intact. How spreading tissues balance flexibility with integrity is not yet well understood. Here, we show that keratin intermediate filaments adapt tissue mechanical resilience to the stresses arising in epithelial tissues during spreading. By analysing the expansion of the enveloping cell layer (EVL) over the yolk cell in zebrafish embryos in vivo, we find that keratin network maturation in EVL cells is promoted by stresses building up within the spreading tissue.
Through genetic interference and tissue rheology experiments, complemented by a vertex model with mechanochemical feedback, we demonstrate that stress-induced keratin network maturation in the EVL increases tissue viscosity, to prevent tissue rupture. Further, keratins are required in the yolk cell for mechanosensitive actomyosin network contraction and flow, the forces pulling the EVL. These dual mechanosensitive functions of keratins enable a balance between pulling force production and EVL mechanical resilience, ensuring uniform and robust tissue spreading.
This repository is organized as described in the tree below. A brief description of the various features and functions is provided as comments.
.
├── README.md # This readme file that explains the basic architecture and function of this repo
├── pixi.toml # pixi configuration file to run the code
├── pixi.lock # Locked dependency versions for reproducible environment installation
├── data/ # Data files for figures 1, 2, 3, 4, and 7
│ └── Fig1/
├── data_vertex_model/ # Vertex-model simulation outputs
│ └── simulation_scripts/ # Vertex-model scripts to launch simulations
├── Figure/ # Exported figure panels
│ └── Figure1/
├── notebooks/ # Notebooks that source the data in the data folder and plot it
│ └── Figure1_KeratinNetworkMaturationEVL.ipynb
├── src/ # Shared plotting/utilities that I think are neat!
│ ├── BinnedPlotter.py # Bin intensity data over time and plot Figure intensity traces
│ ├── Spanselection_pipette.py # A nifty script I made to analyze pipette aspiration data (span selection + fitting)
│ ├── plot_vertex_model.py # Plots all the panels of the vertex model data. Author: Yann-Edwin Keta
│ ├── epiboly_dynamics_archive.py # Plots mean-field predictions of epiboly dynamics. Author: Silke Henkes
│ ├── pipette_dynamics_parvary_archive.py # Plots mean-field predictions of pipette dynamics. Author: Silke Henkes
│ ├── batlow.py # Scientific colormap. Author: Fabio Crameri
│ └── Figureplot.py # Planned figureplot function
└── .gitignore
The notebooks are designed to be used with pixi to manage the environment and run Jupyter notebooks.
-
Install the environment (first time only):
pixi install -
Start JupyterLab from the project root:
pixi run jupyter lab-
or Use the
jupservetask to run JupyterLab without a browser and bind to all interfaces:pixi run jupserve -
or Use the pipette aspiration analyssis script by using:
pixi run python src/Spanselection_pipette.pyFor a deeper tutorial please check out the Note about this script
-
-
Open any .ipynb file and run cells from there to recreate each panel for the figures.
The figures and plots use the following style elements which you would need to have on your system. These colors are kept standard and are colorblind friendly!
- Fonts: Helvetica and Arial
- Color scheme hex ids:
- Main color: #83BB03
- Other green: #27BB03
- Brown (depreciated): #BB9703
- Orange : #ff7f0e
- Blue: '#0383bb'
- Ruby pink: '#bb0383'
Spanselection.py was developed to analyze heightmaps from pipette asipration data and fit the asipration and retract curves automatically. The goal of this function is to ease analysis of pipette aspiration or other such that and provide fits directly in python with a simple gui. It can be run with the command pixi run python src/Spanselection. It opens a window where the same graph is represented two allowing
-
#1 To fit the aspiration curves
-
#2 To fit the retract curves
These badges indicate the main tools, environments, and dependencies used in the analyses:
- Python: core language used across scripts and notebooks.
- Version 3.13: version targeted by the environment.
- Pixi: environment and dependency manager for reproducible setup.
- Jupyter: notebook interface for running analyses.
- PyTorch: deep learning framework used in specific workflows.
- Cellpose: segmentation tool used for image analysis.
- Matlab: For PIV analysis and Deproj based image projections
- Suyash Naik: suyash.naik@ista.ac.at
- Yann-Edwin Keta: yann-edwin.keta@espci.fr
The authors would like to thank Xavier Trepat, Edouard Hannezo and Jiri Friml for discussions and comments on the manuscript. We would like to ackowledge Diana Pinhero, Nicoletta Petridou, Christoph Sommer, Irene Strecarri and present and past members of Heisenberg and Hannezo labs. We also would like to thank the Imaging & Optics, Aquatic facility, Electron Microscopy, and Scicomp Units at the Institute of Science and Technology Austria (ISTA). We would like to thank Makoto Furutani-Seiki and Masazumi Tada for providing transgenic lines. This research was supported by a FWF, grant DOI 10.55776/PAT5044023 and JKU Nanocell grant DOI 10.55776/W1250. We acknowledge support from the EMBO and FWF for travel to conferences.
We would like to make the data and code in this repository available for FAIR use under the CC BY-SA 4.0 status.
Completed, mostly....