A beneficial gene can sweep through a microbial community without sweeping away the genomes around it. This model suggests that patchy habitats, plus only moderate horizontal gene transfer, may be enough to preserve genome diversity. Classic population genetics predicts that an acquired beneficial gene causes a genome-wide selective sweep: the genome spreads clonally with the beneficial gene, removing genome diversity.

Metagenomic data instead show gene-specific sweeps, where a beneficial gene spreads across a bacterial community while maintaining genome diversity. The proposed scenario is a metacommunity where species occupy multiple patches and a realistic, moderate HGT rate helps maintain genome diversity.

In this model, diversity decreases only moderately when a new beneficial gene arrives, and losses can be quickly restored; the key parameters are beneficial-gene insertion frequency, migration rates, and horizontal transfer rates. Horizontal gene transfer plays a crucial role in the processes that shape bacterial evolution.

It accelerates adaptation to new ecological niches and reduces the deleterious effects associated with genetic load from clonal reproduction. HGT is also a widespread pathway through which pathogenic bacteria acquire resistance to antibiotics.

When HGT is low, vertical inheritance is the main mechanism for expanding novel genetic variants, so evolutionary dynamics are described by clonal evolution. Because of linkage effects, a highly beneficial gene can cause a drastic reduction in diversity as the mutant genome clonally expands and sweeps together with the beneficial gene.

Metagenomic evidence supports gene-specific sweeps, where a beneficial gene reaches fixation across species or strains without erasing diversity. Standard population-genetic models would require very high recombination rates to reconcile that pattern, compared with direct measurements of those rates.

That gap suggests that more complex mechanisms should be involved. The proposed complementary mechanism has metacommunity structure as its key ingredient. The environment contains multiple patches, such as nutrient patches in marine snow, and physical separation limits the spread of beneficial genes through genome migration or HGT.

Under this structure, diversity can be preserved during fixation of a beneficial gene without requiring high recombination rates. The model describes community-wide gene-pool diversity in a spatially structured environment that supports different species during gene sweeps.

Each species dominates a single habitat, so intra-habitat dynamics are typically neutral fixation or a selective sweep of one species or strain. The model does not address intra-population diversity; diversity refers only to pan-metagenomic diversity.

Figure one schematically follows a multi-species metacommunity through time: distinct patches contain single, phenotypically homogeneous populations, shown as differently colored shapes. A beneficial gene, marked by a cross, can emerge, spread between patches through horizontal gene transfer, and then shape drift and selection dynamics, including reproduction and death.

The visual matters because it summarizes how migration, selection, and gene exchange jointly determine both species composition and the distribution of beneficial genes. A beneficial gene can spread through migration of an individual and its genome, producing a genome-wide sweep, or through HGT and a gene sweep in the community.

Migration may reduce system-wide diversity by replacing the species in an invaded patch, whereas HGT transfers the beneficial gene across genetic backgrounds without losing diversity. When diversity is maintained and regenerated by new species, the neutral innovation rate creates a time scale of order one over the innovation rate, allowing diversity recovery to compete with genome-wide sweeps.

The model counts time in steps, with one move occurring at the end of each step, and also uses a metacommunity meta-generation in which one generation equals M time steps. The diversity-maintenance process uses neutral migration-substitution events, where one existing species replaces another in a patch.

It also uses innovation events, meaning the emergence of a new species through external migration or speciation. The neutral model provides a benchmark without HGT and contains two elementary events: neutral migration or sweeps across patches, and innovation events corresponding to new species.

At each time step, a patch population can change through an innovation event, such as speciation or neutral migration and sweep from an outside species pool, with rate nu per patch and per time step. Figure two establishes the paper’s neutral benchmark without horizontal gene transfer: patches either receive a new species through innovation or are replaced by a migrating species.

The simulations show diversity relaxing toward a plateau, while panel C compares that equilibrium with the analytical expression across innovation rates. Panel D shows trajectories collapsing when time is scaled by the equilibration scale, supporting an intrinsic timescale proportional to the number of patches divided by the innovation rate.

In the neutral regime, diversity displays equilibration dynamics and reaches a stationary state called S zero. The analytical equilibrium expression is computed in the limit of large M and small innovation rate. Numerical simulations agree with the analytical expression and show that the typical equilibration time is the inverse of the innovation rate.

The next question is how much diversity is lost when a beneficial gene is introduced without diversity-maintenance mechanisms. The beneficial gene can spread through both HGT, producing a gene-specific sweep, and genome-wide sweeps on individual patches. The initial diversity comes from the neutral model, and the analysis assumes that beneficial-gene fixation is much faster than neutral diversity equilibration.

This is the most conservative scenario for the introduction of a beneficial gene because it is the most adverse in terms of diversity loss. Figure three separates two ways the beneficial gene spreads: migration sweeps an entire patch and reduces diversity, whereas horizontal gene transfer spreads the gene while preserving existing species.

In panel B, simulations show diversity falling from its initial level, labeled S sub i, to a stable residual level, S sub f, after fixation. Panels C and D compare the predicted sweep parameter with simulations across horizontal-transfer-to-migration ratios and innovation rates, showing how the model quantifies the remaining diversity.

Under these assumptions, two processes occur: migration-sweep of a patch by a beneficial-gene carrier, which reduces diversity, and HGT spread of the gene, which does not reduce diversity. Diversity can therefore only decrease in this scenario, and the analysis measures the decrease relative to the diversity baseline.

For a genome-wide migration-sweep, two patches are chosen randomly, and if the first carries the beneficial gene while the second does not, the second species is replaced by a copy of the first. For an HGT gene-sweep, two random patches are selected and the beneficial gene transfers into the second patch without displacing its species.

The model uses independent HGT and migration rates on a fully connected network, so spatial distance between patches is not modeled. The number of patches carrying the beneficial gene follows logistic growth, and after the fixation time all species in the metacommunity carry the advantageous gene.

This evolutionary regime is defined by the fixation time being much shorter than the equilibration time. The key quantity is the residual diversity after the beneficial gene fixes. After the beneficial gene is introduced, diversity falls from its initial value to a new stationary value.

The sweep parameter compares the final diversity with the neutral baseline; a value of zero means a genome-wide sweep across the metacommunity, while a positive value means some diversity is regenerated. The analytical prediction agrees with numerical simulations under small innovation rate, small HGT rate, and large metapopulation size.

A full genome sweep occurs only when the HGT-to-migration rate ratio approaches zero, so invasion by the beneficial-gene species is the only relevant process. For any positive HGT-to-migration ratio, more than one species remains after fixation of the beneficial gene.

At an HGT-to-migration ratio of zero point one and innovation rate zero point zero one, the sweep parameter is about zero point five, meaning that an HGT-sweep rate ten times slower than the typical migration-sweep time regenerates half the diversity in the worst-case scenario.

The model next examines longer time scales, where multiple gene sweeps interact with diversity restoration. Three forces act at each time step: innovation, migration of a beneficial-gene species into a patch without the gene, and HGT transfer followed by a gene sweep.

In an innovation event, the new species carries the beneficial gene with probability equal to the fraction of patches already carrying that gene. Neutral migration can also move species with the same genetic content, whether both species carry the beneficial gene or both lack it.

Figure four shows how beneficial-gene sweeps interact with innovation-driven diversity restoration. Panel B separates the initial diversity drop during gene fixation from the later recovery toward the starting level, while panel C shows the minimum scaled diversity parameter increasing with the innovation rate.

With repeated gene emergence, panel D displays oscillations, and panel E shows that maximum diversity changes once the scaled emergence rate passes a critical timescale set by fixation and equilibration. With zero innovation rate, the minimum diversity equals the value obtained without any diversity-restoring mechanism.

With a positive innovation rate, the minimum diversity is always higher than it is without diversity restoration. This happens because the diversity drop and the diversity-restoring mechanism compete on different time scales; if restoration is fast enough, diversity cannot reach its natural minimum.

The simplified metacommunity model supports gene sweeps without eliminating genome diversity. Gene sweeps can produce a moderate reduction in diversity even without a diversity-restoring mechanism, while adding such a mechanism can increase the minimum observed diversity.

High rates of beneficial mutations can instead reduce maximum diversity, and the preservation mechanism is compatible with small HGT rates compared with typical migration time scales. The model’s limitations come from a trade-off with its simplicity, including no spatial organization, no relationship between migration and HGT rates, and a highly simplified treatment of intra-population dynamics.

It includes only neutral non-beneficial mutations, omits deleterious mutations, and does not investigate clonal interference effects. Deleterious mutations could reduce overall genetic diversity and modify sweep time scales through linkage effects. The model outcomes rely on simple time-scale competition arguments, connecting the hypothesis to Hutchinson’s proposal that changing environmental conditions can prevent competitive exclusion from reaching equilibrium.

A proposed experiment would induce gene sweeps in a laboratory metacommunity with different patch densities to test the role of patchiness in restoring diversity. The experiment could compare spatial frequency-spectrum or genetic-diversity patterns with the variance in the number of co-existing species predicted by different models.

The model also predicts that migration affects residual diversity after a gene sweep, which could be tested in laboratory gene-sweep setups. Future genomic studies may distinguish gene-specific sweeps with or without high HGT by analyzing additional selective forces in other genome regions.

The central takeaway is that metacommunity structure can make gene-specific sweeps compatible with moderate HGT: diversity drops only moderately and can recover, offering a testable explanation for genomic diversity without exceptionally high transfer rates.