🔬 condensed matter

Continuous limit of a discrete stochastic model of cell migration

This paper analytically derives the continuous limit of the one-dimensional Cellular Potts Model, revealing that while it accurately describes cell migration under low forces, it exhibits significant algorithmic artifacts at higher forces which are substantially mitigated by replacing the Metropolis update rule with Glauber dynamics.

Original authors: Despeignes Nino, Durand Marc

Published 2026-06-03

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Despeignes Nino, Durand Marc

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a cell not as a tiny, perfect sphere, but as a squishy, stretchy blob of jelly trying to move through a crowded hallway. Scientists often use a computer simulation called the Cellular Potts Model (CPM) to watch these blobs move, grow, and squish around. Think of the CPM as a grid of tiny floor tiles. The cell is a cluster of these tiles, and the computer moves the cell by randomly trying to swap a tile from the cell with a tile from the empty hallway.

However, there's a catch. The computer doesn't move the cell because it "wants" to go somewhere; it moves the cell based on a set of mathematical rules (like a game of chance) designed to minimize energy. This works great for simulations, but it makes it hard for physicists to connect the dots between the computer's "game rules" and the real-world laws of physics (like how a car accelerates when you press the gas pedal).

This paper is like a translator. The authors took that complex computer game and mathematically figured out what the "real" physics looks like when you zoom out and stop looking at the individual tiles.

Here is the breakdown of their findings using simple analogies:

1. The "Low Force" Zone: Smooth Sailing

When the cell is being pushed gently (a "low force"), the computer simulation behaves exactly like a real physical object.

The Analogy: Imagine pushing a heavy box across a floor with a gentle, steady hand. The box moves at a speed proportional to how hard you push, and it wobbles a little bit due to random bumps.
The Finding: In this gentle zone, the authors proved that the messy computer rules simplify into a clean, standard physics equation (called a Langevin equation). This means the simulation is trustworthy here. You can predict exactly how fast the cell will move and how much it will wiggle just by knowing the "push" and the "temperature" of the simulation.

2. The "High Force" Zone: The Computer Glitches

When the authors pushed the cell harder (a "high force"), things got weird. The computer simulation started showing "glitches" that don't exist in real physics.

The Glitch 1 (The Speed Limit): In real life, if you push harder, you go faster. In the simulation, the cell hits a "speed limit" and stops speeding up, even if you push harder. It's like a car with a broken accelerator that maxes out at 30 mph no matter how hard you floor it.
The Glitch 2 (The Magic Diffusion): In real physics, how fast a particle wiggles (diffusion) usually doesn't change just because you push it. In the simulation, the harder they pushed, the more the cell wiggled. The "wobble" became dependent on the "push," which is a fake artifact created by the computer's math, not real biology.
The Glitch 3 (Broken Rules): There is a famous rule in physics (the Einstein relation) that links how fast something moves to how much it wiggles. In the high-force simulation, this rule broke. The math stopped making sense.

3. The Fix: Changing the "Game Rules"

The computer simulation uses a specific rule for deciding whether to move a tile, called the Metropolis rule. The authors tested a different rule called Glauber.

The Analogy: Think of Metropolis as a strict bouncer who only lets you in if you meet a very specific, rigid criteria. Glauber is a more flexible bouncer who uses a smoother, more gradual approach.
The Result: By switching to the Glauber rule, the authors found that the "glitches" (the speed limit and the magic diffusion) happened much later. The simulation stayed "real" and physically accurate over a much wider range of forces. It's like upgrading the car's engine so it can actually go faster before the speed limiter kicks in.

4. The "Run-and-Tumble" Cell

Finally, they looked at cells that don't just move in a straight line but move forward, stop, spin around, and move in a new direction (like bacteria).

The Finding: They figured out how to calculate the total "wobble" (diffusion) of these cells. They found that the total movement is a mix of two things: the natural jiggling of the cell (like a hot potato) and the extra movement caused by the cell changing direction (the "run-and-tumble").
The Twist: They noticed that in the computer model, the "jiggling" and the "turning" were mathematically tangled together, making it hard to tell which was which. They suggested a simple math tweak (renormalizing the force) to untangle them, allowing scientists to clearly see how much movement comes from the cell's internal energy versus its decision to turn.

The Bottom Line

This paper is a bridge. It tells computer scientists and biologists: "If you use the Cellular Potts Model, here is exactly when it behaves like real physics and when it starts lying to you."

They showed that:

Gentle pushes work fine with standard rules.
Hard pushes create fake physics artifacts.
Switching to Glauber rules fixes many of these fake artifacts, making the simulation more reliable for studying how cells actually move.

Essentially, they took a complex, pixelated computer game and wrote the instruction manual for how to interpret the pixels as real, continuous motion.

Technical Summary: Continuous Limit of a Discrete Stochastic Model of Cell Migration

Problem Statement
The Cellular Potts Model (CPM) is a widely used computational framework in computational biology for simulating multicellular systems, offering advantages in handling topological changes (division, death, intercalation) and complex shapes. However, a fundamental gap exists in the theoretical understanding of CPM: its dynamics arise from a stochastic optimization process (minimization of a free energy functional via Monte Carlo updates) rather than a Langevin-type equation of motion that explicitly links forces to velocity. Consequently, deriving the continuous limit of CPM dynamics at the cell scale from the underlying lattice-scale dynamics remains an open problem. This is particularly critical for the active matter community, which seeks to relate simulation parameters to physical quantities like mobility and diffusion coefficients.

Methodology
The authors analytically derive the continuous limit of the CPM for a single cell in a one-dimensional space. The study focuses on two driving scenarios: a constant force (representing chemotaxis or constant motility) and a "run-and-tumble" force (representing active Brownian dynamics).

The methodology involves:

Model Definition: The CPM is defined on a 1D lattice with a Hamiltonian containing an elastic compression term and a driving force term acting on the cell's center of mass.
Master Equation Formulation: The authors formulate master equations for the probability distributions of the cell length ( $L$ ) and the center-of-mass position ( $X$ ).
Coarse-Graining: By performing a Taylor expansion of the master equations, the authors derive Fokker-Planck equations governing the evolution of $L$ and $X$ in the continuous limit.
Dynamics Comparison: Two distinct Monte Carlo update rules are investigated and compared:
- Metropolis dynamics: The conventional update rule in CPM.
- Glauber dynamics: An alternative rule often used in active matter modeling.
Analytic Derivation: The authors derive analytic expressions for the drift velocity ( $v$ ) and diffusion coefficient ( $D$ ) as functions of simulation parameters (lattice size $a$ , compressibility $\kappa$ , temperature $\beta^{-1}$ , and force $F$ ) for both update rules.

Key Contributions and Results

Derivation of Fokker-Planck Equations: The paper successfully derives the Fokker-Planck equations for cell size and position. For Glauber dynamics, the cell length distribution converges to a stationary normal distribution with a variance dependent on the force, effectively softening the cell's compression modulus.
Low-Force Regime Validity: In the limit of low forces ( $\beta Fa \ll 1$ ), the cell's center-of-mass dynamics are accurately described by an overdamped Langevin equation ( $\dot{X} = \mu F + \sqrt{2D}\xi(t)$ ). In this regime, the diffusion coefficient and mobility are independent of the force, and the Einstein relation ( $v/D = \beta F$ ) holds.
Identification of Algorithmic Artifacts: Beyond the low-force regime, the authors identify intrinsic artifacts specific to the CPM algorithm:
- Force-Dependent Diffusion: The diffusion coefficient $D$ becomes dependent on the driving force $F$ .
- Non-Linear Velocity: The relationship between drift velocity and force becomes non-linear and saturates at high forces.
- Breakdown of Einstein Relation: The ratio $v/D$ deviates from $\beta F$ , meaning the system does not satisfy the standard Einstein relation outside the low-force limit.
- Distribution Deviation: The stationary distribution of cell length deviates from the Gibbs-Boltzmann equilibrium distribution due to the coupling between length and position changes in the standard CPM update.
Mitigation via Glauber Dynamics: A significant finding is that replacing the standard Metropolis update with Glauber dynamics significantly mitigates these artifacts. While artifacts still exist at high forces, the dependence of the diffusion coefficient on the force is much less pronounced with Glauber dynamics, thereby broadening the physically valid parameter space for simulations.
Run-and-Tumble Dynamics: For a cell with a tumbling force, the authors derive an effective diffusion coefficient at long times ( $t \gg \lambda^{-1}$ ) given by $D_{eff} = D + v^2/(2\lambda)$ . They note that this effective diffusion term depends on both the effective temperature and the tumbling rate, making it difficult to distinguish between membrane fluctuations and active reorientation without renormalizing the force term in the Hamiltonian.

Significance
The paper claims to bridge the gap between lattice-based simulations (CPM) and continuous active matter models. By providing exact analytic expressions for drift velocity and diffusion coefficients, the work allows researchers to:

Determine the bounds of simulation parameters within which the CPM accurately reproduces continuous Langevin dynamics.
Identify and quantify artifacts that arise when parameters fall outside these bounds.
Justify the use of Glauber dynamics over Metropolis dynamics to reduce algorithmic artifacts and extend the validity of the model.

The authors position this work as an important first step toward understanding the coarse-grained dynamics of the CPM for more complex 2D or 3D multicellular systems, providing a theoretical foundation for interpreting simulation results in terms of physical transport properties.