Dagmar Iber & CoBi

Paper & poster are available on the COMSOL conference website: www.comsol.com/paper/direct...

"Within the competition to bring this field to a new level, SimuCell3D is remarkable and will mark a clear evolution of the topic."
🥁That’s what the committee said about this #SIBRemarkableOutputs 2024
👉Discover the output: tinyurl.com/53arz2rz
@iberd.bsky.social

The goal: make it easier for others to reproduce and extend these models within COMSOL.

We hope this serves as a generalizable reference for simulating collective cell behavior and pattern formation.

arxiv.org/pdf/2509.08930

This complements our earlier work introducing the DCM model 👉 bsky.app/profile/iber...

That previous paper focused on the biological questions and mathematical framework.

Here, we focus on the practical COMSOL #PIDE implementation: setup, BCs, 1D–3D, and Lagrangian reformulation for growth.

Out now: Simulating Organogenesis in #COMSOL: Tissue Patterning with Directed Cell Migration

We provide a detailed walkthrough of how to implement #DCM partial integro-differential equation models - enabling accessible simulations of tissue patterning and morphogenesis.

arxiv.org/pdf/2509.08930

Morphogen gradients can convey position and time in growing tissues Morphogen gradients are known to guide spatial patterning, but can they also encode time? Vetter and Iber propose that a co-expanding morphogen source generates transient signals, allowing cells to measure time without additional molecular clocks. The Sonic Hedgehog gradient in the mouse neural tube is used to show how this mechanism can act as a timer. Opposing gradients synchronize differentiation across the tissue, providing a simple, widely applicable strategy for coordinating space and time during development.

Online now: Morphogen gradients can convey position and time in growing tissues #newton #physics

Morphogen gradients can convey position and time in growing tissues Morphogen gradients are known to guide spatial patterning, but can they also encode time? Vetter and Iber propose that a co-expanding morphogen source generates transient signals, allowing cells to me...

Paper: doi.org/10.1016/j.ne...

Tweetorial: x.com/DagmarIber/s...

Our paper "Morphogen gradients can convey position and time in growing tissues" is now out in Newton ‪@cp-newton.bsky.social‬

Quite fitting to see this novel idea that morphogen gradients not only encode position, but can also time & synchronise development over long distances out in a new journal.

Directed cell migration is a versatile mechanism for rapid developmental pattern formation The evolution of multicellular organisms hinges on self-organization mechanisms that generate tissues with diverse functions. A central process is the breaking of symmetry to form spatial patterns fro...

📄 Read the full study here: doi.org/10.1101/2025...

🎯 Useful for: developmental biology, tissue engineering, pattern formation, and computational modeling

#DevBio #PatternFormation #CellMigration #Morphogenesis #ComputationalBiology

📌 Summary:

• #DCM is a rapid, robust mechanism for developmental pattern formation
• COMSOL FEM implementation makes DCM models numerically accessible
• #DCM patterning parameter ranges and timeframes
• Pattern Orientation via attraction anisotropy or directed tissue growth

👇Thread 🧵(10/11)

2. Dynamic #attraction #zones

Spatially varying cell attraction that changes with tissue growth can guide migrating cells, leading to precise large-scale patterning.

This mimics how tissues form rings, bands, or layered structures in vivo.

👇Thread 🧵(9/11)

We identify two mechanisms for guiding pattern orientation:

1. #Anisotropic #attraction
Cells pulling or migrating more strongly in one direction form aligned stripe-like patterns—e.g., during directional tissue growth.

👇Thread 🧵(8/11)

#DCM naturally leads to unoriented patterns—spots, labyrinths—similar to Turing-like systems.

But biological tissues often require oriented patterns to fulfill specific functions.

Can DCM produce stripes, too?

👇Thread 🧵(7/11)

Three key parameters drive the emergence and morphology of patterns:

• Initial density of motile cells
• Intercellular attraction strength
• Cell sensing radius

👇Thread 🧵(6/11)

Simulations and linear stability analysis allowed us to find #critical #conditions for pattern formation and predict #patterning #speed.

We show under which conditions #DCM can realistically pattern tissues in development.

👇Thread 🧵(5/11)

We developed a mathematical framework that represents a wide range of #DCM cues, e.g., chemotaxis, durotaxis, haptotaxis & a general Finite Element Method #FEM:

👉 1D, 2D, 3D
👉 arbitrary geometries & boundary conditions
👉 isotropic & anisotropic interactions
👉 fast, large-scale simulations

🧵(4/11)

To study #DCM, both discrete and continuum models have been used.

But:
👉 Discrete models are computationally expensive.

👉 Continuum models have required custom Finite Volume Method #FVM implementations—until now.

👇Thread 🧵(3/11)

During embryonic development, cellular tissues transition from uniform starting conditions into robust spatial patterns.

#DCM offers a particularly fast and versatile route to spontaneously symmetry breaks and pattern formation without tissue buckling.

🧵(2/11)

How can cells self-organize rapidly into complex patterns during development?

Let’s explore a powerful and underappreciated mechanism: Directed Cell Migration (DCM).

Preprint @biorxivpreprint.bsky.social : doi.org/10.1101/2025...

👇Thread 🧵(1/11)

Let me add our paper to this review of mechanical drivers of NT folding.

We showed that mammalian spinal #neuraltube folding is primarily driven by #mesoderm expansion and #zippering. #Hingepoints emerge as a result of these extrinsic forces.

www.pnas.org/doi/10.1073/...

x.com/DagmarIber/s...

Epithelial tissues are well-known to invade free space and stop proliferation at confluency.

Do we know what regulates cell proliferation spatially and temporally?

🧵by @dedenonmathieu.bsky.social

PhD Position - 3D Cell-Based Tissue Simulation Framework

Are you a C++ expert and interested in making 3D cell-based tissue simulations with #SimuCell3D ever more powerful? Please apply and join us as #Master student, #PhD student or #Post-Doc!!

jobs.ethz.ch/job/view/JOP...

jobs.ethz.ch/job/view/JOP...

x.com/DagmarIber/s...

SimuCell3D … for simulating tissue mechanics with cell polarization. @iberd.bsky.social

Dimension: 3D
Software language: C++
Download: git.bsse.ethz.ch/iber/Publica...

bsky.app/profile/epim...

iber / Publications / 2025_almanstoetter_pinnverse · GitLab GitLab Community Edition

👩💻 Want to try it? Just a few extra lines of code vs. classic PINNs!
Code & details: git.bsse.ethz.ch/iber/Publica...

Retweet 🔄 if you’d use this for your inverse problems! #OpenScience

⚡ No plateaus here!
Classical PINNs stall. PINNverse keeps improving physics loss ∼ epoch^(-1.4) for Fisher’s equation.
Algebraic decay >> stagnation!

📉 #PINNverse crushes parameter error — even with:
- High noise
- Terrible initial guesses

Stable & accurate where others fail (e.g., Fisher-KPP).

💥 Why is this a breakthrough?

Standard PINNs miss non-convex Pareto fronts → overfit.
#PINNverse captures the entire Pareto front → balances physics + data perfectly.

🔑 The big idea:

Classical PINNs use weighted-sum loss → often fails.
#PINNverse reframes it as constrained optimization → unlocks better solutions.

Small change, huge impact!

📊 How does #PINNverse stack up?

✅ beats Nelder-Mead & classical PINNs
✅ handles noisy data & bad initial guesses
✅ tested on 4 tough benchmarks:
- Kinetic reaction ODE
- FitzHugh–Nagumo
- Fisher–KPP PDE
- Burgers’ PDE

🚀 Introducing #PINNverse — a game-changer for parameter estimation in differential equations! 🧠💡

No forward solves. Better accuracy. Robust to noise.

Preprint: doi.org/10.48550/arX...

#SciComm #MachineLearning #InverseProblems #PINNs