Fixation Probability

When a new genetic mutation arises in a population, what determines its ultimate destiny? Will it disappear without a trace, or will it spread until it becomes a universal feature of the species? This question of a new allele's likelihood of reaching 100% frequency, known as its ​​fixation probability​​, is central to understanding evolution. The fate of any mutation hangs in the balance between two powerful forces: the directional push of natural selection and the random, unpredictable influence of genetic drift. This article unravels the principles that govern this critical process.

The following chapters will guide you through this fundamental concept. In "Principles and Mechanisms," we will explore the core theoretical framework, from the simple lottery of neutral mutations to the decisive duel between selection and drift. We will see how factors like population size and selective advantage dictate an allele's chance of survival. Then, in "Applications and Interdisciplinary Connections," we will discover how these principles are applied as a master key to unlock mysteries across the life sciences, from reading evolutionary history in our DNA to architecting the future of species with synthetic biology.

Imagine a single new word, a mutation in the language of life, appearing in the immense library of a species' DNA. What determines its fate? Will it be a fleeting whisper, lost to the noise of history, or will it one day be on the lips of every member of the species, its meaning etched into their very being? This journey from a single copy to ubiquity is what we call ​​fixation​​. The question of its likelihood, the ​​fixation probability​​, is one of the most fundamental in all of evolutionary biology. It is a story of chance, of necessity, and of the grand, intricate dance between them.

Let's begin with the simplest possible scenario, the roll of a fair die. Consider a new genetic variant that is ​​selectively neutral​​—it offers no advantage and no disadvantage. It's like a new, unheard-of surname appearing in a vast, isolated population. Its survival is purely a matter of luck.

In a simple population of $N$ haploid bacteria, where each individual is just a single set of genes, a new mutation starts as one copy among $N$. What is its chance of one day being the ancestor of the entire population? Think of it this way: in the great lottery of inheritance, every single gene copy present in the current generation has an equal, independent chance of being the "lucky" one that, generations hence, will have its descendants make up the whole population. Since there are $N$ tickets in this lottery, and our new mutation holds just one, its probability of winning—of reaching fixation—is precisely its starting frequency: $1/N$.

For creatures like us, who are ​​diploid​​ (carrying two copies of each gene), the logic is the same, but the numbers change slightly. If a new mutation arises on one chromosome in a single mountain goat in a herd of $N=250$ individuals, the total number of gene copies in the population is $2N$, or 500. Our new allele's initial frequency is therefore a minuscule $1/(2N)$, or $1/500$. And so, its probability of eventual fixation, if it's neutral, is also just $1/500$, or $0.002$.

This simple, beautiful rule—that for a neutral allele, ​​fixation probability equals initial frequency​​—is the bedrock of our understanding. It immediately reveals a staggering truth: the smaller the population, the greater the odds for any new neutral mutation. A new allele in a population of 100 individuals is 100 times more likely to fix by chance than the same allele in a population of 10,000. This is not because the allele is "better" in a small population, but because the lottery is less competitive.

The force driving this lottery is called ​​genetic drift​​. You can picture it as a drunken walk. In each generation, the frequencies of alleles don't get passed on perfectly; they fluctuate due to the randomness of which individuals happen to reproduce and which of their alleles they happen to pass on. In a very large population—a massive city—the law of large numbers smooths things out. The drunkard takes so many tiny, random steps that they barely move from their starting point. An allele at 50% frequency will likely stay very close to 50%.

But in a small population—a tiny island village—the walk is wild and erratic. A few chance births or deaths can cause huge swings in allele frequencies. The drunkard lurches violently from side to side. An allele can, by sheer luck, stumble all the way to 100% frequency (fixation) or 0% (loss) in a surprisingly short time. Genetic drift is the statistical noise of heredity, and its roar is deafening in small populations.

Now, what if the new allele is not neutral? What if it's better? Suppose a mutation in a firefly gene makes its bioluminescent flash a little brighter, giving it a mating advantage. This advantage is quantified by the ​​selection coefficient​​, $s$. If the new allele gives its bearer a $1\%$ advantage, then $s=0.01$.

You might expect a complex formula to describe its fate, one that involves the population size $N$ and the advantage $s$. But the great biologist J.B.S. Haldane gave us a result of stunning simplicity and power. For a new beneficial mutation with a small advantage $s$, its probability of fixation is approximately:

$P_{fix} \approx 2s$

Take a moment to appreciate this. The population size, $N$, has vanished from the equation!. Why? The intuition is that a new beneficial allele faces its greatest peril at the very beginning, when it exists as just one or a few copies. At this stage, it is desperately vulnerable to being snuffed out by the sheer bad luck of genetic drift—its carrier might fail to reproduce for reasons that have nothing to do with the allele. This initial struggle is a battle against random extinction. If the allele can survive this early gauntlet and increase its numbers, its own selective advantage begins to take over, creating a deterministic rise in frequency that is less perturbed by the random noise of a large population. The fate of the allele is essentially sealed in those first few, precarious generations.

So, for our firefly with a $0.2\%$ advantage ($s=0.002$), its chance of taking over the population is about $2 \times 0.002 = 0.4\%$. This is far better than its chance if it were neutral, which might be one in many thousands.

So we have two great evolutionary forces: the random, drunken walk of drift, whose strength is proportional to $1/N_e$ (where $N_e$ is the effective population size), and the directional hand of selection, whose strength is proportional to $s$. The fate of any mutation hangs on the outcome of the duel between them. The crucial quantity that determines the winner is the product $N_e s$.

• ​​When Selection Rules ($|N_e s| \gg 1$):​​ If the population is large and the selective advantage is significant, selection is the undisputed king. The allele's fate is determined by its fitness. A beneficial allele's fixation probability approaches $2s$, while a deleterious one is almost certainly eliminated. Drift is just a minor tremor under the powerful march of selection.

• ​​When Drift Rules ($|N_e s| \ll 1$):​​ If the population is small or the selection is vanishingly weak, drift calls the shots. The allele behaves as if it were neutral. A slightly beneficial mutation in a tiny gecko population of $N_e = 50$ with a tiny advantage of $s=0.001$ has a $P_{fix}$ that is only about $1.1$ times better than a purely neutral allele. The faint whisper of selection is drowned out by the roar of drift.

This duel leads to a startling conclusion. Can a harmful gene become fixed in a population? Natural selection says no. But genetic drift says, "Perhaps!" In a very small founding population, like 12 "Somnolent Marsupials" colonizing an island, the force of drift can be so immense that it can overwhelm weak negative selection. By sheer chance, a slightly deleterious allele ($s = -0.02$) can lurch its way to fixation, even though it makes the population less fit. This is a powerful illustration that evolution does not always lead to perfection; in small populations, it can be a pathway to decline.

The story so far provides a powerful framework, but reality adds fascinating layers of complexity. The $P_{fix} \approx 2s$ formula, while elegant, hides the brutal reality of the initial struggle for survival. A single copy of a new beneficial allele is like a lottery ticket with a $2s$ chance of winning. This is often a very small number. Most beneficial mutations, even those with a strong advantage, are lost to drift within the first few generations.

However, if an allele is lucky enough to survive this initial phase and establish a small foothold—say, it increases to three copies—its prospects brighten considerably. It has survived the first, most dangerous roll of the dice. The probability that all three of its lineages go extinct is much lower than the probability that a single lineage does. Consequently, its fixation probability is now almost three times its initial value. Evolution is a game of survival, and surviving the start is more than half the battle.

Finally, we must remember that genes do not live in isolation. They are passengers on chromosomes, traveling together through the generations. Imagine our brilliant beneficial mutation arises on a chromosome that, by chance, is also carrying a collection of slightly harmful mutations—some "bad luggage." In regions of the genome where ​​recombination​​ (the shuffling of genes between chromosomes) is rare, the beneficial allele is shackled to its deleterious neighbors. Selection cannot simply pick the good gene; it must act on the whole package.

This phenomenon, known as ​​background selection​​, means the beneficial allele is constantly dragged down by its linked, unfit companions. Its effective selection coefficient is reduced. The new mutation's chance of fixation is scaled down by the proportion of "clean" chromosomes in the population that are free of this bad luggage. This is the Hill-Robertson effect: linkage between genes interferes with the efficiency of natural selection. It's a beautiful, deep insight that connects the physical structure of our genome—our chromosomes and their recombination hotspots and deserts—directly to the speed and power of evolution. The fate of a single gene is tied to the fate of its entire neighborhood.

From a simple lottery to the complex interplay of chromosomal geography, the story of fixation probability reveals the heart of the evolutionary process: a delicate, and often unpredictable, dance between the randomness of drift and the directionality of selection.

We have seen that the ultimate fate of a new genetic mutation—whether it vanishes into obscurity or rises to conquer a population—is governed by a subtle interplay between the deterministic push of selection and the chaotic jostling of random chance. This concept of fixation probability is far more than an elegant piece of mathematics; it is a master key that unlocks fundamental processes across the landscape of the life sciences. It allows us to read the deep history written in our genomes, to predict the course of epidemics, to understand the geographic spread of species, and even to take the reins of evolution itself. Let us now explore a few of these fascinating applications.

One of the most profound and beautiful insights from the theory of fixation is its application to neutral evolution. Imagine a mutation that has no effect on an organism's survival or reproduction—a silent change in the genetic code. As we have discussed, its probability of fixation is simply its initial frequency. For a single new mutation in a diploid population of size $N$, this is a tiny number: $1/(2N)$. However, new mutations are constantly arising. If the mutation rate per gene copy is $\mu$, then in every generation, a total of $2N\mu$ new neutral mutations appear in the population.

What, then, is the long-term rate at which one neutral allele is replaced by another? This rate, which we can call the substitution rate $k$, is simply the number of new mutations appearing per generation multiplied by their probability of fixing. The result is astonishingly simple:

$k = (2N\mu) \times \left( \frac{1}{2N} \right) = \mu$

The population size $N$—a term that seems so central to the story—vanishes from the equation completely! This stunning result, a cornerstone of the Neutral Theory of Molecular Evolution, tells us that for neutral parts of the genome, the rate of evolution is equal to the underlying mutation rate. This provides us with a "molecular clock." If we can estimate the mutation rate, we can compare the genetic sequences of two species and calculate how long ago they shared a common ancestor.

This clock, however, does not tick uniformly everywhere in the genome. It runs fastest in regions where mutations are most likely to be neutral. Consider a pseudogene—a gene that has been disabled by mutation and no longer produces a functional product. A new mutation in this genetic fossil is unlikely to have any effect on the organism. In contrast, consider a gene that codes for a vital enzyme. Here, almost any random change will be harmful, or deleterious. Selection will weed out these new mutations with ruthless efficiency. While a neutral mutation in a pseudogene has a fixation probability of $1/(2N)$, a deleterious mutation in an essential gene will have a much, much lower chance of ever reaching fixation. Thus, by comparing the rate of evolution in different parts of the genome, we can distinguish functional, constrained regions from non-functional "junk" DNA. The slow-ticking parts of the genome are where the most important functions are encoded.

The power of fixation probability extends beyond identifying constraint; it allows us to actively hunt for adaptation. A central tool in modern genomics is the ratio of nonsynonymous to synonymous substitution rates, known as $ω$ or $d_N/d_S$. Synonymous mutations change the DNA sequence but not the amino acid sequence of the resulting protein, and are often assumed to be nearly neutral. Nonsynonymous mutations change the amino acid and are thus visible to selection.

By comparing the fixation rates of these two classes of mutations, we can infer the selective pressures on a gene.

• If $\omega 1$, nonsynonymous mutations are fixing less often than neutral ones. This is the signature of ​​purifying selection​​, which weeds out deleterious changes to preserve a protein's function. This is the most common signal seen across genomes.
• If $\omega \approx 1$, nonsynonymous mutations are fixing at roughly the neutral rate. The gene is likely evolving without significant constraint or advantage.
• If $\omega > 1$, nonsynonymous mutations are fixing more often than neutral ones. This is a tell-tale sign of ​​positive selection​​. It suggests that the environment is changing and that new variations in the protein are advantageous, driving rapid adaptation. Finding genes with $\omega > 1$ allows us to pinpoint the genetic basis of evolutionary innovations, like the development of antibiotic resistance in bacteria or the adaptation of a virus to a new host.

But nature is subtle, and not all that glitters is positive selection. There are molecular processes that can mimic its signature. One such imposter is GC-biased gene conversion (gBGC). During the process of genetic recombination, the cellular machinery that repairs DNA sometimes has a slight preference for using G and C bases over A and T bases. This creates a transmission bias, a "push" in favor of G/C alleles that is entirely independent of the organism's fitness. This bias acts like a small, positive selection coefficient for any mutation that increases GC content. It can increase the fixation probability of slightly deleterious nonsynonymous mutations, inflating the $d_N$ rate and creating a false signal of $\omega > 1$. This teaches us a crucial lesson: to truly understand the story in our DNA, we must account for all the forces that can influence the fate of an allele, not just classical natural selection.

The size and structure of a population can have dramatic consequences for the power of genetic drift, and therefore for the fate of mutations. When a population shrinks to a very small size—a phenomenon known as a bottleneck—drift can overpower even moderately strong selection. Imagine a virus jumping from its natural host to humans. The founding viral population in the first human may be established by just a handful of viral particles. In this tiny population, even a slightly deleterious mutation has a surprisingly high chance of fixing purely by luck, a chance many times higher than in the large, stable host population it came from. This principle is critical in conservation genetics, where endangered species in small, fragmented populations can accumulate harmful mutations, and in epidemiology, where it helps explain the unpredictable evolution of new pathogens.

Population structure in space can also lead to surprising outcomes. During a rapid range expansion, new neutral mutations that arise at the leading edge of the expanding wave have a vastly higher probability of reaching a high frequency than mutations that arise in the stable core of the population. This phenomenon, known as "gene surfing," occurs because the founders of each new colony are drawn from the small population at the very edge of the front. A lucky allele that is present in these founders gets to ride the wave of expansion, spreading across a vast geographic area not because it is better, but simply because it was in the right place at the right time. This connects the abstract math of fixation probability to the concrete patterns of biogeography we see in the world today.

For most of history, we have been observers of evolution. But by mastering the principles of fixation probability, we are beginning to become architects. Nature herself provides templates for how to "cheat" the standard rules. For instance, some genes, known as meiotic drivers, are able to ensure they are passed on to more than their fair 50% share of offspring. A neutral mutation that happens to be physically linked to such a driver allele on the same chromosome can "hitchhike" to fixation, swept along by the selfish success of its neighbor. The fate of this neutral allele has nothing to do with its own properties and everything to do with the company it keeps.

Synthetic biologists now use these principles to direct evolution in the laboratory. In Adaptive Laboratory Evolution (ALE) experiments, scientists use serial dilution—growing a culture of microbes and then transferring a small fraction to fresh media—to apply strong selective pressures. The size of this dilution directly sets the severity of the population bottleneck in each cycle. By tuning this factor, researchers can modulate the effective population size, thereby controlling the relative strength of selection and drift. A smaller transfer volume (a more severe bottleneck) makes drift more powerful, while a larger transfer volume allows selection to act more efficiently. This allows them to predictably guide the fixation of beneficial mutations and rapidly engineer organisms with desired traits.

The most powerful and provocative application of these ideas is the CRISPR-based gene drive. A gene drive is a feat of genetic engineering that creates an artificial version of a meiotic driver. An allele is engineered to not only encode a desired trait, but also the machinery to cut the other chromosome and copy itself into the gap. This converts heterozygotes into homozygotes, ensuring the allele is passed to nearly all offspring. This mechanism gives the drive allele an enormous transmission advantage, equivalent to a very strong positive selection coefficient. Even if it provides no fitness benefit to the organism, its fixation probability can be extraordinarily high, approaching 100%. The potential to use these systems to, for example, eliminate mosquito populations that carry malaria or to eradicate invasive species is immense, as are the ethical considerations. It represents a profound new stage in our relationship with the living world, one made possible by a deep understanding of the fundamental principles governing the fate of a single gene.