Evolution is usually taught as a two-step process. Random mutations create genetic variation, and natural selection decides which variants become common. That picture has been enormously productive, but it has never been completely uncontested. For more than fifty years, a minority proposal has argued that organisms might do more than simply wait for useful mutations: they might evolve mechanisms that reduce mutation rates in the genes they can least afford to change. If true, the most important genes would be quieter parts of the genome, while genes where change could be useful would be more active. On the face of it, the idea sounds sensible. Essential genes do vital work, and damage to them can kill a cell. Why would evolution not protect them?
A new study from the University of Bath says the evidence for that protection does not hold up. Dr Sofia Radrizzani and Professor Laurence Hurst analysed what they describe as the largest and cleanest collection of spontaneous mutations assembled to date, drawing on experiments in the bacterium Escherichia coli and baker's yeast. Their tests found no relationship between how essential a gene is and how often it mutates. The work, published in PLOS Biology, does not revolutionise evolutionary theory. Instead, it removes a challenge to the standard view and explains why earlier evidence may have looked convincing when it was not. The distinction between mutation and selection, the authors conclude, remains intact.
A disputed idea with a long history
The hypothesis that important genes mutate less often was first proposed in 1972. It has resurfaced repeatedly since, including a notable renewal of debate in 2023. Its appeal is easy to understand. Genomes are not uniform. Some regions mutate more than others, and some genes are clearly more important to survival than others. It is a short step from those observations to the suggestion that the two patterns might be connected.
The University of Bath announcement describes the claim as a challenge to the traditional view of evolution that has remained contentious for decades. The new paper, titled The importance of genes does not predict their mutation rate, is the most direct test yet. It appeared in PLOS Biology, volume 24, article e3004089, on 12 November 2026, with the DOI 10.1371/journal.pbio.3004089. PubMed indexes it as PMID 43019505, and the full text is available through PubMed Central.
Radrizzani, who co-wrote the paper as part of her PhD, is blunt about the history of the idea. "The claim that mutation is slower for important genes is a triumph of hype over substance," she says. "It's a really old suggestion and is reinvented in various forms routinely." That framing is important. The pair are not merely reporting a negative result; they are arguing that a long-running line of evidence has been misinterpreted.
What it would mean to protect a gene
To understand the claim, it helps to be precise about mutation rates. A mutation rate is the probability that a given piece of DNA changes during replication or repair. Rates vary across genomes for many reasons, including chromosome position, chromatin structure and the timing of replication. The disputed hypothesis adds a functional element: it says that natural selection has shaped those rates so that essential genes are protected.
Under this view, a gene involved in copying DNA or maintaining basic cellular machinery would be expected to mutate slowly. A gene where a change might confer an advantage, such as one affecting antibiotic resistance in bacteria or beak dimensions in birds, might be expected to mutate more quickly. The organism would not be choosing which genes to change. Instead, over many generations, mechanisms that bias mutation rates in this way would spread because they improved survival.
The conventional alternative is simpler. Mutation rates arise from molecular processes and genome architecture. Natural selection then acts on the resulting variation, removing harmful changes and favouring helpful ones. Selection can change how common a mutation is in a population, but it does not necessarily change the underlying probability that the mutation occurs. The Bath study is, in effect, a test of whether the extra, functional layer is needed.
Why the intuition is seductive
The idea that organisms protect their most important genes feels almost self-evident. Mutations in essential genes can disrupt vital processes, cause genetic disease or kill the cell. Given such stakes, a mechanism that lowered mutation rates in those genes would seem clearly advantageous. It is the kind of evolutionary argument that can be made in a single sentence.
There is, however, a quantitative problem. Mutations in individual genes are already rare. Reducing them further would provide only a small additional benefit, because most genes are not mutating often enough to be a constant source of danger. For natural selection to favour a protective system, the benefit would need to outweigh the cost of building and maintaining it. The Bath team argues that the margin is too thin.
The authors illustrate the conventional model with Darwin's finches, which Charles Darwin encountered in the Galápagos Islands in 1835. Differences in beak shape allow the birds to exploit different foods. Random mutation produces variation in beak size and shape, and changing food availability favours some birds over others. Over generations, advantageous traits spread. Selection shapes the outcome, not the rate at which the underlying changes appear.
This example does not disprove the protective hypothesis, but it shows why the standard model already explains adaptation without it. The extra claim needs its own evidence, and that is what the new study set out to examine.
The ghost-mutation problem
One of the study's central explanations for earlier apparent support is what Professor Hurst calls "ghost mutations". Harmful mutations in important genes are more likely to be removed by natural selection. If researchers examine the genetic changes still present in a population, they may see fewer mutations in essential genes. But that absence does not mean fewer mutations occurred. It may mean the damaging ones disappeared before they could be recorded.
"We'd expect natural selection to rapidly remove mutations when they occur in more important genes, because these mutations are more likely to be harmful to the organism," Hurst says. "This can create the illusion that important genes have fewer mutations when, in reality, some mutations have simply disappeared before we have a chance to analyze them."
This is a subtle but powerful objection. It turns the observation inside out. Where supporters of the hypothesis see evidence that important genes are protected from mutation, the Bath team sees evidence that selection has been busy after mutation. The two explanations can produce the same pattern in ordinary population data. Distinguishing them requires a study design in which selection is weakened, so that mutations can be observed close to the moment they arise.
Mutation accumulation in bacteria and yeast
Mutation accumulation experiments are designed to minimise the effects of natural selection. In practice, this means imposing repeated bottlenecks in which a single cell or small number of cells found the next generation. Under such conditions, selection is much less efficient at purging harmful mutations, so changes that would normally vanish can be observed.
This design directly addresses the ghost-mutation issue. If the goal is to learn how often mutations arise in different genes, researchers need to look before selection has removed the inconvenient evidence. Mutation accumulation lines provide something close to that view.
The Bath team then asked a simple question: do genes that matter more mutate less? To answer it, they needed a measure of gene importance. They drew on experiments in which individual genes are removed from organisms. Some genes are indispensable, meaning that cells die without them. Others can be deleted with little or no noticeable effect. That gives an experimental ranking, rather than a subjective judgement about which genes seem important.
The resulting analysis compared this ranking with observed mutation rates across the genome. Under the protective hypothesis, essential genes should consistently appear at the low-mutation end of the distribution. Under the conventional view, they should not.
Checking for hidden bias
A negative result is only convincing if the experiment could have detected a positive one. The Bath team therefore ran three independent tests to check whether selection might still have distorted the mutation accumulation data. Selection can never be eliminated completely, and even a small residual effect could, in principle, create the appearance of fewer mutations in important genes.
None of the three tests revealed evidence of substantial bias. That matters because it closes off the most obvious objection. If the data had been shaped by selection, the lack of a pattern might have been an artefact of the method. Instead, the researchers argue that their dataset is close to a direct record of spontaneous mutation.
The team also considered other ways of measuring gene importance. Beyond deletion experiments, they looked at how much protein each gene produces and how strongly natural selection has acted on the gene over thousands of years. A highly expressed gene might be expected to receive special protection, as might one that is extremely conserved across evolutionary time.
These additional measures were included partly to make the test fair. If the protective hypothesis could survive under any reasonable definition of importance, the researchers wanted to give it the chance. It did not.
Measuring gene importance three ways
The first measure, essentiality, comes from deletion studies. It is the most direct: remove a gene and see whether the cell survives. The second, protein abundance, asks whether genes that produce large quantities of protein are less likely to mutate. The third, evolutionary constraint, asks whether genes that have changed little over long periods show lower mutation rates.
Each measure captures a different intuition about what "important" means. A gene can be essential in the laboratory but not strongly conserved. A gene can be highly expressed without being indispensable. A gene can be conserved for reasons unrelated to mutation rate. By testing all three, the team reduced the chance that a single imperfect metric was driving the conclusion.
The results were consistent. "We found that genes producing more abundant proteins don't have lower mutation rates," Radrizzani says. "Nor does the way genes evolve over thousands of years predict how frequently they mutate." Hurst adds: "Analyze the data any which way and we found nothing. In neither species did these key genes have a lower mutation rate."
The phrase "any which way" is a fair summary of the paper's strategy. Rather than relying on one analysis, the authors tested multiple definitions and multiple statistical approaches. All pointed in the same direction.
What the analysis found
The headline result is absence. Essential genes did not mutate less often than non-essential genes. Genes under strong evolutionary constraint did not mutate less often than genes that evolve quickly. Highly expressed genes did not mutate less often than genes with modest output. In both E. coli and baker's yeast, the proposed relationship simply was not there.
Negative findings can be difficult to communicate because they lack a single dramatic number. But the scale of this dataset gives the result weight. With a large collection of spontaneous mutations from two organisms and several independent definitions of importance, the study had a strong opportunity to detect an effect if one existed.
The authors are careful about what they claim. They do not say that every gene has the same mutation rate. They say that variation in mutation rates is not explained by the importance of the gene. That is a narrower conclusion, but an important one, because it separates molecular causes from functional ones.
The paper therefore supports the conventional distinction between mutation and selection. Mutation supplies variation with biases that arise from genome structure and replication. Selection then removes or promotes variants. What selection does not appear to do, on this evidence, is tune mutation rates gene by gene according to usefulness.
Structure, replication timing and real variation
The study did find patterns, just not the one predicted by the protective hypothesis. Genes copied later during DNA replication tend to accumulate mutations more frequently than genes copied earlier. This is a structural explanation. It reflects the mechanics of replication and the cellular environment in which different parts of the genome are copied.
Other known causes of variation include chromosome position and the chemical environment around DNA. These effects are real and measurable. They can explain why mutation rates differ across a genome without invoking a system that shields vital genes. In other words, the genome is not uniform, but its non-uniformity appears to be mechanical rather than purposeful in the sense proposed by the 1972 hypothesis.
This distinction matters for how researchers interpret mutation data. If a gene shows a low mutation rate, the first explanation to consider should be its genomic context and replication timing, not its biological importance. Conversely, an essential gene that mutates often is not evidence that protection has failed; it may simply sit in a part of the genome where mutations are common.
Hurst summarises the position: "While we see known causes of variation in mutation rates, we found no evidence for the effects that have been proposed as a challenge to the conventional model. None predicted which genes would have high or low mutation rates, and there are certainly no grounds for a shift in paradigm."
What the result does not overturn
It would be easy to overread this study as a claim that mutation is perfectly random. That is not what the authors say. Mutation rates vary, sometimes substantially, and the paper documents some of that variation. The claim is narrower: the variation does not track gene importance in the way the protective hypothesis predicts.
Nor does the result diminish the role of natural selection. Selection remains the process that shapes which mutations persist. The paper actually strengthens selection's explanatory role by arguing that apparent protection was probably selection removing harmful changes after they arose. In that sense, the study is conservative. It keeps the standard framework and removes a proposed addition.
There are also boundaries to the work. The experiments concern two microorganisms, both of which reproduce quickly and have relatively compact genomes. Plants and animals, with larger genomes and more complex development, may present additional complications. The authors are not claiming to have settled every version of the hypothesis for every species. They are reporting a strong failure to replicate the expected pattern in the systems where it should be easiest to see.
Why a negative result matters
Science often celebrates discoveries of new mechanisms. Less attention is paid to studies that show a proposed mechanism is unnecessary. Yet those studies are essential, because they prevent fields from accumulating explanations that are attractive but unsupported. The Bath paper performs that housekeeping role for evolutionary biology.
The 1972 idea kept returning because it addressed a genuine puzzle: harmful mutations are common, and essential genes matter. The answer offered by the new study is that selection itself creates the misleading pattern. Harmful mutations in important genes vanish quickly, leaving behind an apparent quietness that is really the absence of survivors. Recognising that process clarifies rather than complicates evolution.
For students, the result is a useful correction to a common misconception. Natural selection does not need to anticipate which genes are important. It does not plan, and it does not rewrite mutation rates on demand. It acts on the variation that mutation provides, with all the biases that molecular biology imposes. The elegance of evolution lies partly in how much can be explained without adding extra design-like mechanisms.
The paper's conclusion is therefore modest in wording but significant in scope. There is no need to rewrite Darwin's theory. The largest mutation analysis yet assembled found no shield around the genome's most important genes, and no grounds for changing the framework that has guided evolutionary biology for generations. That may not be a dramatic result, but in a field crowded with bold claims, a careful negative result can be the more durable contribution.
Nexuswild welcomes factual corrections. Email contact@nexuswild.com with evidence and the article URL.
