Subdiffusive random growth of bacteria

⚕️

This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine a bacterium as a tiny, living balloon that needs to grow from a small size to a big size before it splits into two. Scientists have long known how these bacteria control their final size (like a strict rulebook for when to split), but they were puzzled by how the balloon actually inflates minute-by-minute. Is the growth smooth and steady, or is it a jittery, chaotic mess?

This paper by Wei and Lin dives into that minute-by-minute chaos and discovers something surprising: the bacteria aren't just growing randomly; they are growing with a specific kind of "memory" that comes from their physical skin, not their brain.

Here is the story of their discovery, broken down into simple concepts:

1. The "Fake" Acceleration (The Statistical Trap)

First, the researchers had to clean up a mess. Previous studies suggested that bacteria speed up their growth right before they divide, like a car flooring the gas pedal at the finish line.

However, Wei and Lin realized this might be an illusion caused by how we look at the data.

The Analogy: Imagine watching a runner on a track. If you only look at runners who are almost at the finish line, you might accidentally pick the ones who are sprinting hard to catch up, while ignoring the ones who are already far ahead and slowing down. This creates a false impression that "everyone speeds up at the end."
The Fix: They built a computer simulation of a "perfectly steady" growing bacterium. When they analyzed this fake data using the same old methods, it also looked like it was speeding up at the end. This proved that the "speed up" was partly a trick of the math, not just biology.

2. The Real Discovery: The "Jelly" Effect (Subdiffusion)

Once they corrected the math, they looked at the remaining "jitter" in the growth. They expected the growth rate to be like white noise—think of it as static on an old TV, where every flicker is totally random and unrelated to the one before it.

Instead, they found subdiffusion.

The Analogy: Imagine you are walking through a crowded, sticky swamp. If you take a step forward, the mud grabs your foot. If you try to step forward again immediately, the mud is still holding you. You have to wait for the mud to let go. Your movement isn't random; it's "sticky." If you move fast now, you are likely to slow down next because of the resistance you just encountered.
The Result: The bacteria's growth rate has a "memory." If it grows a little too fast, it tends to slow down immediately after. If it lags, it tends to catch up. This creates a "sticky" pattern where the growth rate fluctuates much slower than expected. The math showed this "stickiness" follows a specific rule (an exponent of 0.27), which is very different from normal random noise.

3. The Culprit: The Cell Wall (Not the Genes)

Why does this "sticky" growth happen?

The Biological Guess: Maybe the bacteria are turning genes on and off to control growth?
The Rejection: Genes and proteins take minutes or hours to change. The "stickiness" the researchers saw happens in seconds. It's too fast for the cell's "brain" (DNA) to be the cause.
The Physical Answer: It's the cell wall. Bacteria have a tough outer shell made of a mesh called peptidoglycan. The researchers realized this shell acts like a complex sponge or a springy rubber band.
- The Analogy: Think of the cell wall as a giant, tangled net of springs and shock absorbers (like a car suspension). When the inside pressure pushes out, the net stretches. But because the springs are different sizes and tangled, some stretch instantly, while others take a long time to relax. This mix of fast and slow reactions creates the "sticky" growth pattern.

4. The Model: The "General Voigt" Machine

To prove this, they built a mathematical model of the cell wall.

They imagined the wall as a long chain of tiny springs and dampers (shock absorbers) connected in a line.
Crucially, they gave these springs a wide variety of "relaxation times" (some are stiff and fast, some are loose and slow), following a specific mathematical pattern (a power law).
The Result: When they ran this model, it perfectly recreated the "sticky," subdiffusive growth they saw in real bacteria.

The Big Takeaway

This paper tells us that the chaotic, wobbly growth of a bacterium isn't just random noise or a complex genetic program. It is largely a physics problem.

The bacterium is like a balloon being inflated inside a complex, stretchy, and unevenly made net. The way that net stretches and relaxes dictates how the balloon grows. The "rules" of growth are written in the mechanics of the cell wall, not just in the DNA.

In short: Bacteria don't just "decide" to grow; they are physically constrained by their own skin, which acts like a complex, sticky sponge that slows down and speeds up their expansion in a very specific, predictable way.

1. Problem Statement

Understanding bacterial cell size regulation is a fundamental challenge in quantitative biology. While intergenerational size control (e.g., the "adder" model) is well-characterized, the stochastic nature of cell volume growth within a single cell cycle remains elusive.

The Debate: Recent studies suggest bacterial growth is not strictly exponential but exhibits "superexponential" acceleration toward the end of the cell cycle. However, it is unclear whether this acceleration is a genuine biological program or an artifact of statistical analysis methods.
The Gap: The fluctuations of cell volume growth on timescales shorter than the cell cycle (minutes) are poorly understood. Specifically, the nature of the noise driving these fluctuations (white noise vs. correlated noise) and its physical origin (biological regulation vs. mechanical constraints) are unknown.

2. Methodology

The authors employed a multi-faceted approach combining statistical benchmarking, experimental data analysis, and physical modeling.

Minimal Drifted Brownian Model (Control):
- The authors constructed a null model where cell volume $V$ follows a drifted Brownian motion: $dz = kdt + k'dW$, where $z = \ln(V/\bar{V}_b)$ is the log-volume, $k$ is a constant deterministic growth rate, and $dW$ is white noise.
- Cell division was simulated using the "adder" model.
- Purpose: To serve as a controlled benchmark to test common statistical frameworks. Since the ground truth growth rate is constant, any observed variation in the analysis must be a statistical artifact.
Statistical Analysis of Experimental Data:
- Analyzed experimental E. coli data (from Ref. [5]) using three conditioning methods:
  1. Relative Age: Conditioning on the fraction of the cell cycle elapsed ( $\phi$ ).
  2. Relative Volume: Conditioning on the current log-volume ( $z$ ).
  3. Elapsed Time: Conditioning on time since birth.
- Calculated the Mean Square Displacement (MSD) of the log-volume fluctuations, $\sigma^2_z(\Delta t)$ , to characterize the noise dynamics.
Physical Modeling (General Voigt Model):
- Hypothesized that the cell wall (peptidoglycan network) acts as a complex viscoelastic material.
- Modeled the cell wall as a series of $N$ Voigt elements (spring-dashpot pairs) with relaxation times ( $\tau$ ) distributed according to a power law: $P(\tau) \sim \tau^{-\beta}$ .
- Replaced the white noise in the growth equation with "colored noise" generated by this viscoelastic network to simulate the stress fluctuations driving volume growth.

3. Key Contributions

A. Identification and Correction of Statistical Biases

The study rigorously demonstrated that standard statistical conditioning introduces systematic biases that can be mistaken for biological signals:

Relative Age Bias: Even in the minimal model with constant growth, conditioning on relative age $\phi$ creates an apparent "superexponential" acceleration near the end of the cycle ( $\lambda \sim (1-\phi)^{-1/2}$ ). This is a pure mathematical divergence, not a biological phenomenon.
Relative Volume/Time Bias: Conditioning on volume or elapsed time in the minimal model produces an apparent deceleration near the end of the cycle due to trajectory truncation effects.
Conclusion: After accounting for these artifacts, the experimental data still shows a genuine growth-rate acceleration in the second half of the cell cycle, confirming the "superexponential" hypothesis as a real biological feature, distinct from the statistical noise.

B. Discovery of Subdiffusive Growth Dynamics

The most significant finding is the characterization of the stochastic fluctuations in cell volume:

Subdiffusion: The MSD of the log-volume fluctuations scales as $\sigma^2_z(\Delta t) \sim \Delta t^\alpha$ .
Anomalous Exponent: The measured exponent is $\alpha \approx 0.27$ .
Implication: Since $\alpha < 1$ , the growth rate noise is not white noise. It exhibits strong negative temporal correlations (memory) on the timescale of minutes. This means a positive fluctuation in growth rate is likely to be followed by a negative one, and vice versa, suppressing the variance over time.

C. Mechanical Origin of Fluctuations

The authors argue that this subdiffusion arises from physical mechanics rather than biological regulation:

Timescale Argument: The fluctuations occur on a minute timescale, much faster than gene expression or protein synthesis (which operate on doubling times).
Viscoelastic Model: The General Voigt model successfully recapitulates the observed subdiffusive behavior.
Parameter Mapping: The model links the macroscopic growth exponent $\alpha$ to the microscopic material exponent $\beta$ of the peptidoglycan network via the relation $\alpha = 3 - 2\beta$ .
Result: Fitting the data yields $\beta \approx 1.37$ , suggesting the cell wall is a heterogeneous material with a broad distribution of relaxation times, fundamentally constraining growth stochasticity.

4. Results

Validation of Superexponential Growth: By separating statistical artifacts from biological signals, the study confirms that E. coli growth accelerates in the late cell cycle.
Quantification of Noise: The growth rate noise is subdiffusive with $\alpha \approx 0.27$ .
Model Success: The General Voigt model, treating the cell wall as a power-law distributed viscoelastic network, perfectly reproduces the experimental MSD scaling and the value of $\alpha$ .
Heterogeneity: The results imply that the peptidoglycan meshwork is not a uniform plastic shell but a complex material where stochastic bond cleavage and strand insertion create mechanical heterogeneity.

5. Significance

Paradigm Shift in Growth Modeling: The paper challenges the standard assumption that bacterial growth noise is white (uncorrelated). It establishes that growth is governed by subdiffusive dynamics with long-range memory.
Physical vs. Biological Regulation: It suggests that short-timescale growth fluctuations are governed by physical constraints (cell wall mechanics) rather than active biological regulatory programs. This shifts the focus of size control research toward the biophysics of the cell envelope.
Methodological Rigor: The work provides a critical cautionary note for the field of single-cell biology. It demonstrates that naive statistical conditioning can generate false positives (e.g., apparent acceleration) and emphasizes the necessity of using minimal null models to validate biological hypotheses.
Material Science Connection: It bridges microbiology and soft matter physics, showing that the power-law relaxation of the bacterial cell wall dictates the statistical properties of cell growth.