Purifying selection is the most common form of natural selection, and it works by removing harmful mutations from a population over time. Rather than driving dramatic evolutionary change, it acts as a quality control system, keeping genes and other functional DNA sequences working the way they should. Every generation, new mutations arise, and purifying selection is the force that filters out the ones that would make an organism less likely to survive or reproduce.
How Purifying Selection Works
Every time DNA is copied and passed from parent to child, mistakes slip in. In humans, each newborn carries roughly 66 new single-nucleotide mutations compared to their parents’ genomes. Most of these land in stretches of DNA that don’t do much, so they have no real effect. But some fall in genes or regulatory regions where the existing sequence matters. If a mutation makes a protein fold incorrectly, disrupts a critical regulatory signal, or otherwise reduces an organism’s ability to survive and reproduce, that individual is less likely to pass the mutation on. Over generations, these harmful variants get weeded out. That’s purifying selection.
The process doesn’t require anything dramatic. An organism carrying a mildly harmful mutation might simply have slightly fewer offspring, or its offspring might be slightly less healthy. Multiply that small disadvantage across hundreds or thousands of generations, and the mutation gradually disappears from the population. Strongly harmful mutations get eliminated quickly, sometimes in a single generation if they prevent reproduction entirely. Mildly harmful ones linger longer but still trend toward removal.
The Scale of Its Influence
Purifying selection shapes far more of the genome than most people realize. One analysis estimated that it influences as much as 85% of the human genome, with biased patterns of DNA transmission affecting most of the remainder. Less than 5% of the human genome appears to evolve purely by chance. Cross-species comparisons between humans, mice, and macaques suggest that 5 to 6.5% of nucleotide sites are under direct functional constraint, meaning mutations at those positions are actively selected against.
The mutation load in humans illustrates the scale of the problem purifying selection solves. With about 66 new mutations per person per generation and roughly 5.7% of the non-repetitive genome under selective constraint, each person inherits an estimated two or more new harmful mutations from their parents. If all these mutations acted independently on survival, the math gets startling: models predict that the vast majority of individuals in each generation would need to fail to reproduce just to keep the population’s overall fitness stable. In reality, mutations interact with each other in complex ways, and sexual reproduction helps shuffle them into combinations that selection can act on more efficiently.
Why Population Size Matters
Purifying selection works best in large populations. In a big population, a mildly harmful mutation reduces its carrier’s success relative to many competitors, so the mutation is reliably filtered out over time. In a small population, random chance plays a bigger role. A person carrying a slightly damaging mutation might get lucky and reproduce anyway, simply because there aren’t enough individuals for the statistical disadvantage to matter. This means small populations tend to accumulate more mildly harmful variants than large ones.
Population bottlenecks, where a group suddenly shrinks in size, can weaken purifying selection’s grip. Isolated human populations that went through historical bottlenecks show signs of this: more harmful variants reaching higher frequencies, and more stretches of the genome that are identical between individuals. However, strongly damaging mutations still get removed even in reduced populations. The most harmful variants are under such intense pressure that purifying selection remains effective regardless of population size. It’s the mildly harmful ones that slip through the cracks.
Purifying Selection vs. Other Types
Natural selection comes in several flavors, and the easiest way to distinguish them is by what they do to genetic variation. Purifying selection reduces it. By removing harmful variants and preserving the sequences that already work well, it pushes populations toward genetic uniformity at functional sites. Positive selection also reduces variation, but for a different reason: it drives a new, beneficial mutation to high frequency, sweeping out alternatives along the way. Balancing selection does the opposite of both, actively maintaining multiple versions of a gene in the population. Under balancing selection, rare variants have an advantage precisely because they’re rare, which prevents any single version from taking over.
Of these three, purifying selection is by far the most pervasive. Positive selection gets more attention because it produces the visible adaptations that make evolution exciting to talk about. But the genome’s day-to-day evolutionary work is mostly custodial: holding the line against the constant influx of harmful mutations.
How Scientists Measure It
The standard tool for detecting purifying selection is a ratio that compares two types of mutations in protein-coding genes. Some mutations change the amino acid a gene produces (these can affect function), while others are “silent,” swapping one DNA letter for another without altering the protein. Scientists compare the rate of functional changes to the rate of silent changes. If functional changes accumulate more slowly than silent ones, the ratio falls below 1.0, signaling that purifying selection is removing the functional mutations before they can stick. The lower the ratio, the stronger the purifying selection. A ratio of exactly 1.0 means no selection at all, and above 1.0 signals positive selection favoring change.
Beyond coding regions, scientists detect purifying selection by looking at patterns of variation across populations. Regions under purifying selection show fewer differences between species than you’d expect by chance, and within a population, the harmful variants that do exist tend to be rare. An excess of low-frequency variants at a given genomic position is a hallmark of purifying selection pushing against mutations that keep arising but never gaining a foothold.
Genes Under Extreme Purifying Selection
Some genes are so critical that purifying selection has kept them nearly identical across vast evolutionary distances. Histone H4, a protein that helps package DNA inside cells, is one of the most conserved proteins known. Its amino acid sequence is virtually identical across all plants, animals, and fungi, despite these lineages diverging hundreds of millions of years ago. The DNA sequences encoding histone H4 do vary between species (silent mutations accumulate freely), but the protein itself stays the same because any change to its structure would disrupt DNA packaging in every cell. Studies have confirmed that this extreme conservation is driven by strong purifying selection rather than any unusual property of the DNA itself.
Histones aren’t unique in this regard. Genes involved in fundamental cellular processes, like copying DNA, producing energy, or building ribosomes, tend to show very low rates of functional change across species. These are the genes where almost any mutation is a bad one, so purifying selection is relentless.
Non-Coding DNA Under Selection
Purifying selection doesn’t only protect genes that encode proteins. Large stretches of non-coding DNA, regions that don’t produce proteins but regulate when and where genes turn on, are also conserved. These conserved non-coding sequences often function as switches that control gene activity during development or in specific tissues. In fruit flies, an estimated 85% of sites within these conserved non-coding regions are under selective constraint, with the strength of selection at each site being substantial enough to reliably remove new mutations.
This finding was important because it ruled out an alternative explanation: that these regions simply mutate less often. If they were just “cold spots” for mutation, you’d expect a uniform reduction in variation. Instead, the pattern matches purifying selection specifically, with harmful variants appearing but staying rare, exactly as you’d see if organisms carrying mutations in these regions were less fit. The same pattern appears in mammals, confirming that purifying selection maintains regulatory DNA across very different branches of the animal kingdom.