Transition from traveling fronts to diffusion-limited growth in expanding populations

1. Problem Statement

The paper addresses a discrepancy between theoretical models of microbial colony expansion and experimental observations.

• Theoretical Context: Standard reaction-diffusion models (e.g., Fisher-Kolmogorov-Petrovsky-Piskunov or FKPP equations) predict that microbial colonies expand as traveling fronts with a constant velocity ($v$), implying the colony radius ($R$) grows linearly with time ($R \propto t$). This assumes that biomass redistribution (motility) is independent of nutrient concentration.
• Experimental Reality: Many experiments on non-motile, dense cellular aggregates (bacteria, yeast) show that while some colonies grow linearly, others exhibit sublinear growth where the radius scales as the square root of time ($R \propto t^{1/2}$). Consequently, the colony area grows linearly, not the radius.
• The Gap: Previous models incorporating nutrient limitation (diffusion-limited growth) still predicted linear radius growth ($R \propto t$) because they assumed biomass motility was constant or independent of local nutrient levels. The authors hypothesize that the missing link is the dependence of biomass redistribution (motility) on nutrient concentration. Experimental evidence suggests that mechanical rearrangements and cell movement require active growth, which is fueled by nutrients.

2. Methodology

The authors developed a modified reaction-diffusion model and analyzed it using both analytical derivations and numerical simulations.

• Model Formulation:
  They modified the standard nutrient-limited model (Eq. 3) to couple biomass motility directly to nutrient concentration. The new system of equations (Eq. 4) is:
  $$\frac{\partial b}{\partial t} = D_b \nabla \cdot (bn \nabla b) + \gamma bn$$
  $$\frac{\partial n}{\partial t} = D_n \nabla^2 n - \gamma bn$$
  Where:

  • $b$: Biomass density.
  • $n$: Nutrient concentration.
  • $D_b, D_n$: Diffusion constants for biomass and nutrients.
  • $\gamma$: Nutrient consumption rate.
  • Key Innovation: The diffusion term for biomass is $D_b \nabla \cdot (bn \nabla b)$. This implies the effective diffusion coefficient is proportional to $n$. If nutrients are depleted, biomass cannot move/redistribute.

• Dimensionless Analysis:
  The model was nondimensionalized to isolate a single control parameter, the dimensionless biomass diffusivity $D$:
  $$D = \frac{D_b n_0^2}{D_n}$$
  where $n_0$ is the initial nutrient concentration.

• Analytical and Numerical Techniques:

  • Traveling Front Analysis: Used the "shooting method" and asymptotic matching to solve for traveling wave solutions ($b(x-vt)$).
  • Diffusion-Limited Analysis: For the sublinear regime, they approximated the biomass profile as a moving Heaviside step function and solved the nutrient diffusion equation with an absorbing boundary condition at the colony edge.
  • Simulations: Finite difference methods (explicit for biomass, Crank-Nicolson for nutrients) were used in 1D (narrow channels) and 2D (Petri dish surface) geometries.

3. Key Contributions

1. Novel Mechanism: Identified that the dependence of biomass motility on nutrient concentration is the critical factor distinguishing linear vs. square-root growth regimes.
2. Analytical Solution for Two Regimes: Provided exact analytical solutions for both the linear traveling front regime and the sublinear diffusion-limited regime, including the transition point between them.
3. Unified Framework: Demonstrated that both linear ($R \propto t$) and sublinear ($R \propto t^{1/2}$) growth can emerge from the same physical model depending solely on the magnitude of the biomass diffusivity parameter $D$.
4. Derivation of Scaling Laws: Derived explicit scaling relationships for colony expansion rates, colony height (thickness), and total biomass accumulation as functions of $D$ and spatial dimension $d$.

4. Key Results

A. The Transition Regime

The system exhibits a phase transition at a critical value $D_c \approx 1$:

• High Diffusivity ($D > D_c$): The colony expands as a traveling front.
  • Radius growth: Linear ($R \propto t$).
  • Velocity: $v \approx \sqrt{D} - 1/\sqrt{D}$.
  • Profile: Both biomass and nutrient profiles move as a coherent wave.
• Low Diffusivity ($D < D_c$): The colony expands via diffusion-limited growth.
  • Radius growth: Sublinear ($R \propto \sqrt{t}$).
  • Profile: The biomass profile remains a sharp, invariant step function (moving front), but the nutrient profile does not travel; instead, the nutrient depletion zone widens as $\sqrt{t}$, and the nutrient concentration at the edge decays as $t^{-1/2}$.

B. Colony Morphology and Height ($H$)

The model predicts a relationship between the expansion speed and the colony's vertical thickness ($H$):

• As $D \to 0$ (slow expansion), the colony becomes very thick ($H \to \infty$) because biomass accumulates vertically rather than spreading horizontally.
• As $D \to D_c$, the colony height approaches the initial nutrient concentration ($H \to 1$).
• Analytical relation: $H = 1/\sqrt{D}$.
• Physical Interpretation: In the square-root regime, the colony height is determined by the ratio of nutrient diffusion to biomass motility ($H \propto \sqrt{D_n/D_b}$), independent of the initial nutrient concentration $n_0$.

C. Dimensionality Effects

• 1D (Channels): Expansion rate $\kappa \propto n_0$. Total biomass $B(t) \propto \sqrt{t}$.
• 2D (Circular Colonies): Expansion rate $\kappa \propto \sqrt{n_0}$ (with logarithmic corrections). Total biomass $B(t) \propto t$.
• The geometry dictates how nutrient availability scales with the expansion rate.

D. Validation

• Numerical simulations confirmed the analytical predictions for velocity, colony height, and the coefficient of determination ($R^2$) for linear vs. square-root fits.
• The model successfully reproduces the "frozen" sector boundaries observed in experiments (where motility stops in nutrient-depleted regions) and the "active" boundaries at the edge.

5. Significance

• Resolving Experimental Contradictions: The paper explains why some microbial colonies show linear radius growth while others show linear area growth (square-root radius growth). The difference is not due to different organisms or experimental artifacts, but rather the relative magnitude of biomass motility versus nutrient diffusion ($D$).
• Ecological and Evolutionary Implications: The transition suggests that selective pressure on motility ($D_b$) varies with the environment. In low-nutrient or high-friction environments (low $D$), motility provides little advantage for expansion speed, whereas near the transition ($D \approx D_c$), small increases in motility drastically increase expansion rates.
• Predictive Power: The model provides a minimal set of parameters to predict colony thickness, expansion rate, and nutrient profiles, offering a testable framework for future experiments on biofilms, tumors, and microbial mats.
• Theoretical Advancement: It extends the classic reaction-diffusion paradigm by rigorously incorporating the coupling between mechanical stress (driven by growth) and nutrient availability, moving beyond simple random-walk assumptions for cell movement.