Introduction
In evolutionary biology, the Moran process serves as a fundamental framework for modeling how a single mutant, characterized by fitness $r$, propagates through a population of residents with fitness 1. A recurring question in this field is whether the spatial or social structure of a population—represented as a graph—can act as an "amplifier of natural selection." An amplifier is a structure that increases the probability of a beneficial mutant reaching fixation (taking over the entire population) compared to a well-mixed population.
While researchers have long identified numerous amplifiers under Birth-death (Bd) updating rules, a similar search for amplifiers under death-Birth (dB) updating has consistently come up empty. This research provides a definitive mathematical "no-go" theorem, proving that strong, universal amplifiers for the dB process cannot exist.
The Mechanics of Updating
The distinction between Bd and dB updating is subtle but has profound consequences for evolutionary outcomes.
- Birth-death (Bd): An individual is first chosen to reproduce with probability proportional to fitness. Their offspring then replaces a neighbor.
- death-Birth (dB): An individual is first chosen to die at random. Neighbors then compete to fill the vacancy, with the probability of success proportional to their fitness.
In the dB process, even a highly fit mutant can be randomly selected to die before it has the opportunity to reproduce. This inherent risk creates a "ceiling" on the effectiveness of any population structure to favor selection.
The Impossibility of Strong Amplification
Using Markov chain analysis and rigorous mathematical proofs, the authors establish two primary constraints on dB updating:
1. Boundedness
The fixation probability under dB updating ($\rho^{dB}$) is strictly bounded by $1 - 1/(r+1)$. Unlike Bd processes, where certain structures (like the Star graph) can drive fixation probability toward 1 as population size $N$ increases, dB processes are capped. This ensures that even the most advantageous mutants remain susceptible to loss through random drift.
2. Transience
The research introduces the concept of "transient amplification." While some graphs may appear to amplify selection for small fitness advantages (e.g., $r=1.1$), the authors prove that for any non-complete graph, there exists a threshold fitness $r^*$ (where $r^* \le 2N^2$) beyond which the graph inevitably becomes a suppressor of selection. Thus, amplification in dB is not a universal property of the graph but a temporary phenomenon limited to low-fitness regimes.
Robustness and Sensitivity
The study also explores hybrid models, termed $\delta$-updating, where a population follows dB rules with probability $\delta$ and Bd rules with probability $1-\delta$. The results demonstrate that "super-amplifiers"—structures that guarantee fixation in the limit of large populations—are remarkably fragile. The introduction of even a infinitesimal amount of death-Birth updating (any $\delta > 0$) causes the super-amplification effect to vanish.
While most structures fail quickly, the authors noted that Bipartite graphs exhibit more resilience than Star graphs, maintaining some transient benefit even when the process is 50% death-Birth. Additionally, a newly identified weighted structure called the Fan graph was found to act as a transient amplifier, though it remains unknown if any unweighted, simple graphs can achieve this under pure dB conditions.
Conclusion
This work clarifies a long-standing mystery in evolutionary graph theory by proving that the order of operations—death before birth—fundamentally limits the power of natural selection. It highlights that amplification is not merely a property of a graph's topology, but an emergent feature of the interaction between that topology and the specific biological update rule. Future research may now focus on determining which update rules most accurately reflect specific biological niches, such as stem cell clusters or social networks, to better predict evolutionary trajectories.
Summary updated