Can deleterious mutations surf deterministic population waves? A functional law of large numbers for a spatial model of Muller's ratchet

This paper establishes a functional law of large numbers for a spatial model of Muller's ratchet, proving that the system converges to an infinite system of partial differential equations that rigorously determine population spreading speeds and demonstrate that deleterious mutations can indeed surf deterministic population waves.

The Gist

Imagine a vast, expanding city built by a single family of clones. This family is growing outward, claiming new territory every day. But there's a catch: every time a new child is born, there's a small chance they inherit a "glitch" in their DNA—a deleterious mutation. These glitches don't kill the family immediately, but they make the individuals slightly weaker, slower to reproduce, and less fit for survival.

This paper asks a fascinating question: As this family expands into the empty wilderness, do these "weak" individuals get left behind, or do they ride the wave of expansion all the way to the front?

In biology, this phenomenon is called "gene surfing." It's like a surfer catching a massive wave; if you happen to be at the very front of the crowd when the wave breaks, you get carried far ahead, even if you aren't the strongest swimmer.

Here is a simple breakdown of what the authors discovered, using everyday analogies:

1. The Setup: The "Muller's Ratchet" City

The authors are studying a model called the Spatial Muller's Ratchet.

• The City: Imagine a line of towns (demes) stretching out. People live in these towns.
• The Glitches: Every time someone has a baby, there's a tiny chance the baby gets a new "glitch" (mutation). Once you have a glitch, you can't lose it. You can only get more of them.
• The Ratchet: Because you can't lose glitches, the population's average fitness slowly drags downward over time, like a ratchet wrench that only turns one way.
• The Expansion: The population is pushing into empty land. The people at the very edge (the "front") are the ones colonizing new territory.

2. The Big Question: Can the "Weak" Surf?

In the past, scientists knew that neutral mutations (glitches that do nothing) could surf. If a neutral glitch appears at the front, it can spread wildly just by luck.

But what about deleterious mutations (the bad ones)?

• Intuition: You'd think the weak individuals would be left behind. The strong ones at the front should reproduce faster, pushing the weak ones back.
• The Surprise: The authors wanted to know if the "weak" could somehow hitch a ride on the expansion wave and stay at the front, dragging the whole population's fitness down with them.

3. The Method: From Chaos to Order

The real world is messy. People are born, die, move, and mutate randomly. It's like trying to predict the path of a single grain of sand in a hurricane.

To solve this, the authors used a mathematical trick called a Hydrodynamic Limit.

• The Analogy: Imagine looking at a crowd of people from a helicopter. You can't see individual faces, but you can see the "density" of the crowd. You can see where the crowd is thick and where it is thin.
• The Math: They took their complex, random particle model (the grain of sand) and showed that as the population gets huge, it smooths out into a predictable wave described by equations (PDEs). It's like turning a chaotic storm of raindrops into a smooth, flowing river.

4. The Results: The "Pushed" vs. "Pulled" Wave

The paper confirms some predictions and answers the surfing question with a definitive "No" (but with a twist).

The Spreading Speed:
The speed at which the population expands depends on how much the individuals cooperate or compete.

• Pulled Waves: If competition is high (everyone fights for resources), the wave is "pulled" by the very front. The fastest individuals at the edge lead the charge.
• Pushed Waves: If cooperation is high (people help each other), the wave is "pushed" from behind. The whole crowd moves together.

The Surfing Verdict:
The authors used a clever technique called Tracer Dynamics. Imagine painting a group of people at the start (those who already have mutations) red, and everyone else blue.

• They tracked the red people as the wave moved.
• The Discovery: The red people (the ones with mutations) do not surf. They do not ride the wave to the new frontier.
• Why? The mutations that appear at the front are new mutations. The old, "surfing" mutations get left behind in the "bulk" (the middle of the crowd).
• The Twist: Even though the old mutations don't surf, the front of the wave still ends up full of mutations. Why? Because as the wave moves, new mutations are constantly being born at the front. The front is a factory of new glitches, not a carrier of old ones.

5. The Takeaway

Think of it like a relay race where the runners are getting tired (accumulating mutations).

• Old Mutations: The tired runners from the start of the race get left behind. They don't surf to the finish line.
• New Mutations: However, as the race progresses, the runners at the front start getting tired right there. So, the finish line is always full of tired runners, but they are tired because they just got tired, not because they carried the fatigue from the start.

In simple terms: Deleterious mutations cannot "surf" the population wave in the way we hoped. They don't hitch a ride from the past. Instead, the population expands, and the "bad" mutations are constantly generated fresh at the leading edge, keeping the population's fitness low, but not because of the old ones surfing along.

This rigorous math confirms that while the population expands, the "bad genes" are constantly being re-invented at the front, rather than being carried there by the lucky few from the past.

Technical Summary

Here is a detailed technical summary of the paper "Can deleterious mutations surf deterministic population waves? A functional law of large numbers for a spatial model of Muller's ratchet" by Madeira, Ortgiese, and Penington.

1. Problem Statement and Motivation

The paper addresses the spatial Muller's ratchet, a model describing the accumulation of deleterious mutations in an expanding asexual population. The core biological question is whether deleterious mutations can "surf" (i.e., propagate and become prevalent) on the leading edge of a population wave expanding into an empty habitat.

• Context: In non-spatial models, Muller's ratchet describes the irreversible accumulation of deleterious mutations in asexual populations. In spatial models, "gene surfing" typically refers to neutral mutations becoming fixed at the wave front due to low population density and drift.
• The Gap: While neutral gene surfing is well-studied, the behavior of deleterious mutations in spatially expanding populations remains less understood. Previous work by Foutel-Rodier and Etheridge [34] proposed a spatial particle system model and derived non-rigorous predictions (via generator calculations) regarding the system's hydrodynamic limit (a system of PDEs), spreading speeds, and the potential for deleterious mutations to surf.
• The Challenge: Rigorously proving the convergence of such a system is difficult because:
  1. The system involves infinitely many particle types (particles with $k$ mutations for all $k \in \mathbb{N}_0$).
  2. Birth and death rates are unbounded (polynomial in local density).
  3. There are no a priori bounds on the number of particles per deme.
  4. The interactions are non-local in type space (rates depend on the total local density, coupling all mutation classes).

2. Methodology

The authors employ a multi-step rigorous probabilistic and analytical approach to establish the Functional Law of Large Numbers (FLLN) for the spatial Muller's ratchet.

A. Model Definition

The model is an interacting particle system on a rescaled lattice $L_N^{-1}\mathbb{Z}$.

• State: $\eta^N_k(t, x)$ is the number of particles at deme $x$ with $k$ mutations.
• Dynamics:
  • Migration: Symmetric random walk at rate $m_N$.
  • Birth/Death: Rates depend on local density $\|\eta^N(t, x)\|_{\ell_1}$. Birth rate includes a fitness penalty $(1-s)^k$ and a mutation probability $\mu$ (offspring gets $k+1$ mutations with prob $\mu$).
  • Scaling: $L_N \sim N$ and $m_N/L_N^2 \to m$. The density is rescaled by $1/N$.

B. Proof of Convergence (The Hydrodynamic Limit)

The main technical hurdle is proving that the rescaled density process $u^N$ converges to a solution of an infinite system of PDEs.

1. Tightness in Measure Space: They first establish tightness in the space of Radon measures $D([0, \infty), \mathcal{M}(\mathbb{R})^{\mathbb{N}_0})$ using Jakubowski's tightness criterion. This involves:
   • Proving moment bounds (using results from their companion article [56]).
   • Establishing compact containment conditions.
   • Verifying Aldous' criterion for tightness of evaluations.
2. Tightness in $L^r$ Spaces: To handle the non-linearity and ensure the limit has a density, they prove tightness in a weighted Bochner space $L^{4 \deg q_-}([0, T] \times \mathbb{R}, \hat{\lambda}; \ell_1)$. This relies on:
   • Green's Function Representation: Expressing the density as a martingale plus a drift term involving the reaction term.
   • Regularity Estimates: Using the Green's function to bound space and time increments of the process.
   • Tail Control: Proving that the mass of particles with very high mutation counts vanishes uniformly.
3. Characterization of the Limit: Using Skorokhod's representation theorem, they show any subsequential limit satisfies the integral (mild) formulation of the PDE system.
4. Uniqueness and Regularity: They prove uniqueness of the mild solution in the $L^r$ space using Grönwall's inequality and local Lipschitz properties of the reaction term, establishing that the limit is a unique, smooth (in space and time) solution.

C. Asymptotic Analysis

Once the PDE limit is established, the authors analyze its long-term behavior:

• Mutation Ratios: They derive bounds on the ratio $u_k(t, x) / u_0(t, x)$ using Feynman-Kac formulas and induction arguments.
• Spreading Speed: They adapt arguments from non-local Fisher-KPP equations to determine the asymptotic spreading speed $c^*$.
• Tracer Dynamics: To answer the "surfing" question, they introduce a tracer process $u^*$, representing a subpopulation of initially labelled mutants. They analyze the PDE governing $u^*$ to see if the labelled fraction propagates with the wave.

3. Key Contributions and Results

1. Rigorous Hydrodynamic Limit (Theorem 2.1)

The paper provides the first rigorous proof that the spatial Muller's ratchet particle system converges to the infinite system of PDEs:
$$\partial_t u_k = \frac{m}{2} \Delta u_k + F_k(u)$$
where $F_k$ describes the reaction kinetics (birth, death, mutation). This confirms the non-rigorous conjectures of Foutel-Rodier and Etheridge [34].

2. Bounds on Mutation Prevalence (Theorem 2.5)

In the monostable regime (where the reaction term allows for a stable equilibrium), the authors prove that the ratio of the density of particles with $k$ mutations to those with 0 mutations is uniformly bounded in space and time. Specifically, as $T \to \infty$, the ratio converges to a distribution determined by the mutation-selection balance, bounded by factors involving $Q_{min}$ and $Q_{max}$ (min/max birth rates).

3. Spreading Speed (Theorem 2.7)

For Fisher-KPP type dynamics (where competition dominates cooperation at low densities), the authors rigorously determine the spreading speed $c^*$:
$$c^* = \sqrt{2m((1-\mu)q_+(0) - q_-(0))}$$
They prove that the population front propagates at this speed, and crucially, that all mutation classes (including those with deleterious mutations) spread at this same speed.

4. The Answer to Gene Surfing (Theorem 2.9)

This is the paper's central biological conclusion. By analyzing the tracer dynamics (where only the initial mutant subpopulation is labelled), they prove:

• If the initial population has no labelled individuals without mutations ($f^*_0 \equiv 0$), the density of labelled individuals $u^*(t, x)$ converges uniformly to zero as $t \to \infty$.
• Interpretation: Deleterious mutations cannot surf deterministic population waves. While deleterious mutations are present at the wave front, they are not "surfing" in the sense of being carried by the wave front from the initial mutant pool. Instead, the mutations found at the front are newly generated via mutation-selection equilibrium from the unmutated (fit) population expanding behind them. The "surfing" phenomenon observed in simulations is an artifact of the mutation-selection balance, not the propagation of pre-existing deleterious alleles.

4. Significance

• Mathematical Rigor: The paper overcomes significant technical barriers in the theory of interacting particle systems, specifically handling infinite type spaces and unbounded rates without a priori density bounds. The development of tightness criteria in weighted $L^r$ spaces for $\ell_1$-valued processes is a novel contribution to probability theory.
• Biological Insight: It resolves a debate regarding the mechanism of deleterious mutation accumulation in expanding populations. The result clarifies that in deterministic waves, deleterious mutations do not "surf" (persist and spread because of their initial location); rather, the wave front is constantly "re-mutated" by the high-fitness individuals trailing behind. This distinguishes the mechanism from neutral gene surfing.
• Validation of Models: It validates the use of the specific PDE system derived in [34] for studying spatial evolutionary dynamics, providing a solid theoretical foundation for future simulations and biological applications (e.g., cancer metastasis, invasive species, microbial range expansions).

In summary, the paper bridges the gap between stochastic particle models and deterministic PDEs for complex evolutionary systems and provides a definitive negative answer to the question of whether deleterious mutations can surf deterministic waves, attributing the observed patterns to mutation-selection equilibrium rather than surfing.