Gene Swapping, Not Mutation, Drives the Hidden Evolutionary Rulebook of Gut Bacteria

Every species of gut bacterium carries a hidden record of the evolutionary forces that shaped it, and for decades scientists have read that record with a simple yardstick known as dN/dS. The ratio compares the rate of mutations that change a protein’s amino acid sequence with those that leave it untouched. When purifying selection weeds out damaging changes, the ratio falls far below one; when selection is relaxed, it climbs toward neutrality. A new study of global human gut bacterial populations, published in PLOS Biology by Zhiru Liu and Benjamin H. Good, now shows that this familiar yardstick has been giving researchers a distorted reading, and that the distortion comes from a source most models have ignored: the constant swapping of DNA between strains.

The puzzle at the heart of the study is a broad empirical pattern that has been noted across bacterial genomes for years. When researchers compare pairs of bacterial strains that are only slightly diverged from one another, the measured dN/dS tends to be relatively high. As the comparison moves to more distantly related pairs, the ratio steadily declines. This downward trend has been reported in many bacterial lineages, yet the mechanisms responsible for it have remained contested. Some explanations invoke changes in the strength of natural selection over evolutionary time, while others point to saturation effects or the peculiar demographics of bacterial populations. What most existing models share is an assumption that the genetic differences being measured arise from de novo mutations, the fresh errors that accumulate as lineages split apart and drift away from each other.

Liu and Good argue that this assumption misses a major contributor to bacterial diversity. Bacteria do not simply inherit their genomes clonally from parent cells. Through horizontal gene transfer and homologous recombination, they continually import segments of DNA from related strains, replacing chunks of their own chromosomes with foreign versions of the same genes. In dense microbial communities such as the human gut, this traffic of genetic material can be extensive, meaning that a substantial fraction of the differences between any two strains may not be the product of new mutations at all but of older variants that were acquired from the wider population. Those older variants have had more time to be filtered by selection, and that simple fact, the researchers show, can reshape the entire pattern of dN/dS across a genome.

To disentangle these effects, the team built a phenomenological model of dN/dS in recombining bacterial populations. The model is designed to separate the contributions of two kinds of genomic regions: segments inherited clonally, which trace back through a straightforward line of descent, and segments acquired through recombination from other strains. By applying this framework to pairs of closely related strains, the researchers could ask a question that mutation-only models cannot address: does the mode of inheritance of a DNA segment change the selective signal it carries? The answer, according to their analysis, is a clear yes, and the difference is large enough to account for one of the most widespread patterns in bacterial genomics.

The central finding is strikingly consistent. Clonally inherited regions of the genome, those stretches of DNA passed down directly through the lineage without being swapped, exhibit consistently higher dN/dS ratios than recombined segments. The interpretation follows directly from the age of the variants involved. A freshly arisen mutation in a clonal segment has, on average, only existed briefly, so weakly deleterious changes have not yet been purged and still contribute to the nonsynonymous substitution count. A recombined segment, by contrast, carries variants that circulated in the population before being imported, giving selection a longer opportunity to remove the harmful ones. The result is that recombined DNA looks more constrained, even if the underlying selection pressures on the genes themselves have not changed at all.

Crucially, the researchers did not stop at a qualitative observation. They quantified how the accumulation of recombined segments over time affects the measured ratio, and found that this process can quantitatively explain the majority of the decline in dN/dS with divergence. In other words, much of the trend that has been interpreted as a signal of changing selection pressures across evolutionary time may instead be a bookkeeping artifact: as pairs of strains become more diverged, a growing share of their differences comes from recombined, pre-filtered DNA, dragging the apparent ratio downward. This reframing matters because dN/dS is used throughout microbiology to infer which genes are under selection, to identify adaptive changes, and to estimate the effective strength of natural selection in natural populations. If recombination systematically biases the metric, many of those inferences need to be re-examined.

That is precisely what Liu and Good did next. Armed with a model that accounts for recombination, they revisited the leading explanations for the dN/dS decline in human gut bacteria, including models of purifying selection and adaptive reversion. Adaptive reversion refers to a scenario in which slightly deleterious mutations that fix in small populations are later reversed by compensatory changes, a process that can also depress dN/dS in comparisons between diverged lineages. By fitting their framework to population genomic data from gut bacterial species sampled around the world, the team was able to weigh these competing explanations against the recombination-driven account and to extract new estimates of the selection pressures acting on protein-coding sites.

The re-analysis uncovered evidence for widespread weak selection operating at a large fraction of protein-coding sites in gut bacterial genomes. Weak selection occupies an awkward middle ground in evolutionary genetics: changes under it are not neutral enough to drift freely, yet not strong enough to be efficiently eliminated, particularly in populations whose effective sizes fluctuate. Detecting such selection is notoriously difficult, because its signature is subtle and easily confounded by demographic noise. The finding that a large share of coding sites in gut bacteria experience this regime suggests that the vast majority of the proteome is under some degree of selective constraint, even at positions that a naive analysis might classify as effectively neutral. It also implies that the classical categories of neutral, nearly neutral, and strongly selected sites may need recalibration in species where recombination is rampant.

The broader implications reach beyond the human gut. Horizontal gene transfer is a defining feature of bacterial evolution, shaping everything from antibiotic resistance to metabolic innovation, and recombination among close relatives occurs in many microbial communities, from ocean plankton to soil. If the exchange of older genetic variants systematically alters genome-wide measures of selection, then the standard toolkit of molecular evolution may require revision for any recombining species. The study’s authors frame their conclusions as raising new questions about the effectiveness of natural selection in complex bacterial populations, where the interplay of clonal inheritance, gene flow, and weak selection creates patterns that neither mutation-only nor recombination-free models can capture. For a field that has long treated the bacterial chromosome as a record of mutational history, the message is that the record has been repeatedly edited, and learning to read the edits is now essential for understanding what selection has actually done.

Subject of Research: The role of homologous recombination in shaping genome-wide patterns of selective constraint in human gut bacteria

Article Title: Recombination shapes the dynamics of selective constraint in global populations of human gut bacteria

Article References: Liu, Z., & Good, B. H. (2026). Recombination shapes the dynamics of selective constraint in global populations of human gut bacteria. PLOS Biology, 24(10), e3004052. https://doi.org/10.1371/journal.pbio.3004052

Image Credits: AI Generated

DOI: 10.1371/journal.pbio.3004052

Keywords: dN/dS, recombination, horizontal gene transfer, human gut microbiome, purifying selection, bacterial evolution, population genetics, selective constraint, molecular evolution, PLOS Biology, weak selection, genomic divergence