foambryo 0.2.8

Tension inference for 3D cell assemblies

foambryo

foambryo is a python package based on Polyscope designed to infer surface tensions, pressures and stresses from the geometry of epithelial-cell-clusters, namely early-embryos and structures encoutered in stem cell cultures, as organoids and embryoids.

Living structures encountered in the field of developmental biology have intricate shapes and structures, that reflectes their physiological functions. They embody important features: symmetries, polarisations, patterning. Mechanics is at the heart of the development of these structures, yet tools to investigate forces at play during development in 3D are scarse. Characterizing and quantifying precisely their shapes allows us to match them to a simple physical model, and to reveal the forces shaping cells.

We rely on delaunay-watershed, a tool we designed to reconstruct efficiently multimaterial meshes from instance segmentations. From these multimaterial meshes, we can efficiently and robustly extract junctional angles and surfaces curvatures, and invert Young-Dupré and Laplace laws, to find back the physical parameters involved in the mechanical equilibrium: surface tensions $\gamma_{ij}$ and cell pressures $p_i$.

The viewer is based on Polyscope, a popular viewer designed to visualise 3-dimensional geometrical structures.

Example

Load an instance segmentation, reconstruct its multimaterial mesh, infer and visualize the forces with Polyscope

pip install foambryo

from dw3d import DCEL_Data, geometry_reconstruction_3d
from foambryo import plot_force_inference, plot_tension_inference

## Load the labels
import skimage.io as io
Segmentation = io.imread("Segmentation.tif")

## Reconstruct a multimaterial mesh from the labels
DW = geometry_reconstruction_3d(Segmentation,min_dist = 5)
Verts, Faces_multimaterial = DW.return_mesh()
Mesh = DCEL_Data(Verts,Faces_multimaterial)

## Infer and view the forces
plot_force_inference(Mesh)

#Or just the tensions
plot_tension_inference(Mesh)

Installation

`pip install foambryo

Physical model

We consider a tissue constituted of cells i.

They minimize, under conservation of volume an energy $\mathcal{E} = \underset{ij}{\sum}\gamma_{ij}$ The two main laws involved are:

• Young-Dupré Law: $\vec{\gamma}{ij} + \vec{\gamma}{ik} + \vec{\gamma}_{jk} = \vec{0}$
• Laplace Law: $P_j - P_i = 2 \gamma_{ij} h_{ij}$ where $h_{ij}$ is the mean curvature of the interface between the cell i and j. `

---

API and Documentation

1 - Creating a Mesh object

The first step is to load your multimaterial mesh into a DCEL_Data object via the builder DCEL_Data(Verts, Faces_multimaterial). - Verts is an V x 3 numpy array of vertex positions - Faces_multimaterial is an F x 5 numpy array of face and material indices, where at each row the 3 first indices refers to a vertex and the 2 last refer to a given material, 0 being the exterior media

2 - Infer the forces

Then the second step is to use this Mesh object to infer the tensions and pressions

• infer_tension(Mesh,mean_tension=1,mode='YD'): We infer tensions ratios by inverting junctional equilibrium relations

  • Mesh is a DCEL_Data object
  • mean_tension has to be defined as we only infer ratio between tensions
  • mode is the formula used to infer the tensions. It has to be choosen among: YD, Eq, Projection_YD, cotan, inv_cotan, Lamy, inv_Lamy, Lamy_Log, Variational

• infer_pressures(Mesh,dict_tensions,mode='Variational', P0 = 0): We infer relative pressures by inverting membrane equilibrium relation

  • Mesh is a DCEL_Data object
  • dict_tensions is the dictionnary obtained with infer_tension
  • P0 has to be defined as we only infer relative pressures
  • mode is the formula used to infer the pressures. It has to be choosen among: Variational, Laplace

3 - Visualize

This package is then built around several functions, each of them taking as an entry a Mesh object:

• plot_tension_inference(Mesh,dict_tensions=None,alpha = 0.05, scalar_quantities = False, scattered = False, scatter_coeff=0.2): Which plots surface tensions by inverting Young-Dupré relations

  • Mesh is a DCEL_Data object
  • dict_tensions is the dictionnary obtained with infer_tension, and is computed automatically if unspecified.
  • alpha : p_value threshold when displaying values: Values beyond the alpha and 1-alpha quantiles are clipped
  • scalar_quantities: plot of vertex volume and area derivatives, of mean, gaussian curvatures, and principal curvatures discrepancy. Can be quite long for big meshes
  • scattered: scattered view of the mesh
  • scatter_coeff: amount of displacement if scattered is activated

• plot_force_inference(Mesh,dict_tensions = None, dict_pressure = None,alpha = 0.05, scalar_quantities = False, scattered = False, scatter_coeff=0.2): Which plots surface tensions, pressures and principal directions of the stress tensor

  • Mesh is a DCEL_Data object
  • dict_tensions is the dictionnary obtained with infer_tension, and is computed automatically if unspecified.
  • dict_pressure is the dictionnary obtained with infer_pressure, and is computed automatically if unspecified.
  • alpha : p_value threshold when displaying values: Values beyond the alpha and 1-alpha quantiles are clipped
  • scalar_quantities: plot of vertex volume and area derivatives, of mean, gaussian curvatures, and principal curvatures discrepancy. Can be quite long for big meshes
  • scattered: scattered view of the mesh
  • scatter_coeff: amount of displacement if scattered is activated

• plot_valid_junctions(Mesh): Valid junctions are plotted in green, and unstable junctions are plotted in red. This is used to assess the validity of the inference

  • Mesh is a DCEL_Data object

---

Biological use-cases

Phallusia mammillata

Phallusia mammillata is a solitary marine tunicate of the ascidian class known for its stereotypical development. As the embryo develops freely, without any constraint, we can do a full force inference and infer its tensions, pressures and stresses. We use data from Guignard, L., Fiúza, U. et al.

C.elegans, constrained within a shell

Caenorhabditis elegans is a widely studied model organism, with one of the most reproducible development. This earthworm is developing in a shell. As the shell topology is unknown, the pressures are not accessible. However we can still use Young-Dupré relationships to deduce the surface tensions of the membranes. We use data from the Cao, J., Guan, G., Ho, V.W.S. et al.

View scalar quantities on surface meshes

Gaussian and mean curvature can be plotted on our meshes, and are useful to study the properties of surfaces obtained from simulations. We also plot the vertex area and volume derivatives, that appears in our variational formulas, the difference between the two principal curvatures and the sphere-fit residual that can be used to detect robust non-spherical CMC surfaces (see our paper for more details) They can be obtained by putting the option scalar_quantities = True when viewing the forces.

• Gaussian Curvature is computed using the angle defect formula.
• Mean Curvature is computed using the cotan formula.

To see the code of each of these use-cases, please load the associated jupyter notebooks in the folder Notebooks

---

Credits, contact, citations

If you use this tool, please cite the associated preprint: Do not hesitate to contact Sacha Ichbiah and Hervé Turlier for practical questions and applications. We hope that foambryo could help biologists and physicists to shed light on the mechanical aspects of early development.

@article {Ichbiah2023.04.12.536641,
	author = {Sacha Ichbiah and Fabrice Delbary and Alex McDougall and R{\'e}mi Dumollard and Herv{\'e} Turlier},
	title = {Embryo mechanics cartography: inference of 3D force atlases from fluorescence microscopy},
	elocation-id = {2023.04.12.536641},
	year = {2023},
	doi = {10.1101/2023.04.12.536641},
	publisher = {Cold Spring Harbor Laboratory},
	abstract = {The morphogenesis of tissues and embryos results from a tight interplay between gene expression, biochemical signaling and mechanics. Although sequencing methods allow the generation of cell-resolved spatio-temporal maps of gene expression in developing tissues, creating similar maps of cell mechanics in 3D has remained a real challenge. Exploiting the foam-like geometry of cells in embryos, we propose a robust end-to-end computational method to infer spatiotemporal atlases of cellular forces from fluorescence microscopy images of cell membranes. Our method generates precise 3D meshes of cell geometry and successively predicts relative cell surface tensions and pressures in the tissue. We validate it with 3D active foam simulations, study its noise sensitivity, and prove its biological relevance in mouse, ascidian and C. elegans embryos. 3D inference allows us to recover mechanical features identified previously, but also predicts new ones, unveiling potential new insights on the spatiotemporal regulation of cell mechanics in early embryos. Our code is freely available and paves the way for unraveling the unknown mechanochemical feedbacks that control embryo and tissue morphogenesis.Competing Interest StatementThe authors have declared no competing interest.},
	URL = {https://www.biorxiv.org/content/early/2023/04/13/2023.04.12.536641},
	eprint = {https://www.biorxiv.org/content/early/2023/04/13/2023.04.12.536641.full.pdf},
	journal = {bioRxiv}
}