RNA virus evolution via a fitness-space model

A Model of Epigenetic Inheritance Accounts for Unexpected Adaptation to Unforeseen Challenges

Accumulated evidence of transgenerational inheritance of epigenetic and symbiotic changes raises fundamental questions about the possible types, significance and duration of impacts on the population, as well as whether, and under which conditions, the inheritance of non-genetic changes confers long-term advantage to the population. To address these questions, a population epigenetics model of individuals undergoing stochastic changes and/or induced responses that are transmitted to the offspringis introduced. Potentially adaptive and maladaptive responses are represented, respectively, by environmentally driven changes that reduce and increase the selective pressure. Analytic solutions in a simplified case of populations that are exposed to either periodic or progressively deteriorating environments shows that acquisition and transmission of non-genetic changes that alleviate the selective pressure confer long-term advantage and may facilitate escape from extinction. Systematic analysis of outcomes as a function of population properties further identifies a non-traditional regime of adaptation mediated by stochastic changes that are rapidly acquired within a lifetime. Contrasting model predictions with experimental findings shows that inheritance of dynamically acquired changes enables rapid adaptation to unforeseen challenges and can account for population dynamics that is either unexpected or beyond the scope of traditional models.

Eco-evolutionary dynamics of adapting pathogens and host immunity

As pathogens spread in a population of hosts, immunity is built up, and the pool of susceptible individuals are depleted. This generates selective pressure, to which many human RNA viruses, such as influenza virus or SARS-CoV-2, respond with rapid antigenic evolution and frequent emergence of immune evasive variants. However, the host's immune systems adapt, and older immune responses wane, such that escape variants only enjoy a growth advantage for a limited time. If variant growth dynamics and reshaping of host-immunity operate on comparable time scales, viral adaptation is determined by eco-evolutionary interactions that are not captured by models of rapid evolution in a fixed environment. Here, we use a Susceptible/Infected model to describe the interaction between an evolving viral population in a dynamic but immunologically diverse host population. We show that depending on strain cross-immunity, heterogeneity of the host population, and durability of immune responses, escape variants initially grow exponentially, but lose their growth advantage before reaching high frequencies. Their subsequent dynamics follows an anomalous random walk determined by future escape variants and results in variant trajectories that are unpredictable. This model can explain the apparent contradiction between the clearly adaptive nature of antigenic evolution and the quasi-neutral dynamics of high-frequency variants observed for influenza viruses.

Two fitness inference schemes compared using allele frequencies from 1068 391 sequences sampled in the UK during the COVID-19 pandemic

Throughout the course of the SARS-CoV-2 pandemic, genetic variation has contributed to the spread and persistence of the virus. For example, various mutations have allowed SARS-CoV-2 to escape antibody neutralization or to bind more strongly to the receptors that it uses to enter human cells. Here, we compared two methods that estimate the fitness effects of viral mutations using the abundant sequence data gathered over the course of the pandemic. Both approaches are grounded in population genetics theory but with different assumptions. One approach, tQLE, features an epistatic fitness landscape and assumes that alleles are nearly in linkage equilibrium. Another approach, MPL, assumes a simple, additive fitness landscape, but allows for any level of correlation between alleles. We characterized differences in the distributions of fitness values inferred by each approach and in the ranks of fitness values that they assign to sequences across time. We find that in a large fraction of weeks the two methods are in good agreement as to their top-ranked sequences, i.e. as to which sequences observed that week are most fit. We also find that agreement between the ranking of sequences varies with genetic unimodality in the population in a given week.

Limits to selection on standing variation in an asexual population

We consider how a population of N haploid individuals responds to directional selection on standing variation, with no new variation from recombination or mutation. Individuals have trait values z1,…,zN, which are drawn from a distribution ψ; the fitness of individual i is proportional to [Formula: see text] . For illustration, we consider the Laplace and Gaussian distributions, which are parametrised only by the variance V0, and show that for large N, there is a scaling limit which depends on a single parameter NV0. When selection is weak relative to drift (NV0≪1), the variance decreases exponentially at rate 1/N, and the expected ultimate gain in log fitness (scaled by V0), is just NV0, which is the same as Robertson's (1960) prediction for a sexual population. In contrast, when selection is strong relative to drift (NV0≫1), the ultimate gain can be found by approximating the establishment of alleles by a branching process in which each allele competes independently with the population mean and the fittest allele to establish is certain to fix. Then, if the probability of survival to time t∼1/V0 of an allele with value z is P(z), with mean P¯, the winning allele is the fittest of NP¯ survivors drawn from a distribution ψP/P¯. The expected ultimate change is ∼2log(1.15NV0) for a Gaussian distribution, and ∼-12log0.36NV0-log-log0.36NV0 for a Laplace distribution. This approach also predicts the variability of the process, and its dynamics; we show that in the strong selection regime, the expected genetic variance decreases as ∼t-3 at large times. We discuss how these results may be related to selection on standing variation that is spread along a linear chromosome.

Evolutionary stability of antigenically escaping viruses

Antigenic variation is the main immune escape mechanism for RNA viruses like influenza or SARS-CoV-2. While high mutation rates promote antigenic escape, they also induce large mutational loads and reduced fitness. It remains unclear how this cost-benefit trade-off selects the mutation rate of viruses. Using a traveling wave model for the coevolution of viruses and host immune systems in a finite population, we investigate how immunity affects the evolution of the mutation rate and other nonantigenic traits, such as virulence. We first show that the nature of the wave depends on how cross-reactive immune systems are, reconciling previous approaches. The immune-virus system behaves like a Fisher wave at low cross-reactivities, and like a fitness wave at high cross-reactivities. These regimes predict different outcomes for the evolution of nonantigenic traits. At low cross-reactivities, the evolutionarily stable strategy is to maximize the speed of the wave, implying a higher mutation rate and increased virulence. At large cross-reactivities, where our estimates place H3N2 influenza, the stable strategy is to increase the basic reproductive number, keeping the mutation rate to a minimum and virulence low.

Proliferating active matter

The fascinating patterns of collective motion created by autonomously driven particles have fuelled active-matter research for over two decades. So far, theoretical active-matter research has often focused on systems with a fixed number of particles. This constraint imposes strict limitations on what behaviours can and cannot emerge. However, a hallmark of life is the breaking of local cell number conservation by replication and death. Birth and death processes must be taken into account, for example, to predict the growth and evolution of a microbial biofilm, the expansion of a tumour, or the development from a fertilized egg into an embryo and beyond. In this Perspective, we argue that unique features emerge in these systems because proliferation represents a distinct form of activity: not only do the proliferating entities consume and dissipate energy, they also inject biomass and degrees of freedom capable of further self-proliferation, leading to myriad dynamic scenarios. Despite this complexity, a growing number of studies document common collective phenomena in various proliferating soft-matter systems. This generality leads us to propose proliferation as another direction of active-matter physics, worthy of a dedicated search for new dynamical universality classes. Conceptual challenges abound, from identifying control parameters and understanding large fluctuations and nonlinear feedback mechanisms to exploring the dynamics and limits of information flow in self-replicating systems. We believe that, by extending the rich conceptual framework developed for conventional active matter to proliferating active matter, researchers can have a profound impact on quantitative biology and reveal fascinating emergent physics along the way.

Virus Evolution on Fitness Landscapes

The landscape paradigm is revisited in the light of evolution in simple systems. A brief overview of different classes of fitness landscapes is followed by a more detailed discussion of the RNA model, which is currently the only evolutionary model that allows for a comprehensive molecular analysis of a fitness landscape. Neutral networks of genotypes are indispensable for the success of evolution. Important insights into the evolutionary mechanism are gained by considering the topology of sequence and shape spaces. The dynamic concept of molecular quasispecies is viewed in the light of the landscape paradigm. The distribution of fitness values in state space is mirrored by the population structures of mutant distributions. Two classes of thresholds for replication error or mutations are important: (i) the-conventional-genotypic error threshold, which separates ordered replication from random drift on neutral networks, and (ii) a phenotypic error threshold above which the molecular phenotype is lost. Empirical landscapes are reviewed and finally, the implications of the landscape concept for virus evolution are discussed.

Exploring the expanse between theoretical questions and experimental approaches in the modern study of evolvability

Despite several decades of computational and experimental work across many systems, evolvability remains on the periphery with regards to its status as a widely accepted and regularly applied theoretical concept. Here we propose that its marginal status is partly a result of large gaps between the diverse but disconnected theoretical treatments of evolvability and the relatively narrower range of studies that have tested it empirically. To make this case, we draw on a range of examples-from experimental evolution in microbes, to molecular evolution in proteins-where attempts have been made to mend this disconnect. We highlight some examples of progress that has been made and point to areas where synthesis and translation of existing theory can lead to further progress in the still-new field of empirical measurements of evolvability.

Phenomenological Approach to Cancer Cell Persistence

Drug persistence is a phenomenon by which a small percentage of cancer cells survive the presentation of targeted therapy by transitioning to a quiescent state. Eventually some of these persister cells can transition back to an active growing state and give rise to resistant tumors. Here we introduce a quantitative genetics approach to drug-exposed populations of cancer cells in order to interpret recent experimental data regarding inheritance of persister probability. Our results indicate that alternating periods of drug treatment and drug removal may not be an effective strategy for eliminating persisters.

Population genetics of polymorphism and divergence in rapidly evolving populations

In rapidly evolving populations, numerous beneficial and deleterious mutations can arise and segregate within a population at the same time. In this regime, evolutionary dynamics cannot be analyzed using traditional population genetic approaches that assume that sites evolve independently. Instead, the dynamics of many loci must be analyzed simultaneously. Recent work has made progress by first analyzing the fitness variation within a population, and then studying how individual lineages interact with this traveling fitness wave. However, these "traveling wave" models have previously been restricted to extreme cases where selection on individual mutations is either much faster or much slower than the typical coalescent timescale Tc. In this work, we show how the traveling wave framework can be extended to intermediate regimes in which the scaled fitness effects of mutations (Tcs) are neither large nor small compared to one. This enables us to describe the dynamics of populations subject to a wide range of fitness effects, and in particular, in cases where it is not immediately clear which mutations are most important in shaping the dynamics and statistics of genetic diversity. We use this approach to derive new expressions for the fixation probabilities and site frequency spectra of mutations as a function of their scaled fitness effects, along with related results for the coalescent timescale Tc and the rate of adaptation or Muller's ratchet. We find that competition between linked mutations can have a dramatic impact on the proportions of neutral and selected polymorphisms, which is not simply summarized by the scaled selection coefficient Tcs. We conclude by discussing the implications of these results for population genetic inferences.

Probabilistic analysis of replicator–mutator equations

This paper discusses a general class of replicator–mutator equations on a multidimensional fitness space. We establish a novel probabilistic representation of weak solutions of the equation by using the theory of Fokker–Planck–Kolmogorov (FPK) equations and a martingale extraction approach. We provide examples with closed-form probabilistic solutions for different fitness functions considered in the existing literature. We also construct a particle system and prove a general convergence result to the unique solution of the FPK equation associated with the extended replicator–mutator equation with respect to a Wasserstein-like metric adapted to our probabilistic framework.

Inferring the intrinsic mutational fitness landscape of influenzalike evolving antigens from temporally ordered sequence data

There still are no effective long-term protective vaccines against viruses that continuously evolve under immune pressure such as seasonal influenza, which has caused, and can cause, devastating epidemics in the human population. To find such a broadly protective immunization strategy, it is useful to know how easily the virus can escape via mutation from specific antibody responses. This information is encoded in the fitness landscape of the viral proteins (i.e., knowledge of the viral fitness as a function of sequence). Here we present a computational method to infer the intrinsic mutational fitness landscape of influenzalike evolving antigens from yearly sequence data. We test inference performance with computer-generated sequence data that are based on stochastic simulations mimicking basic features of immune-driven viral evolution. Although the numerically simulated model does create a phylogeny based on the allowed mutations, the inference scheme does not use this information. This provides a contrast to other methods that rely on reconstruction of phylogenetic trees. Our method just needs a sufficient number of samples over multiple years. With our method, we are able to infer single as well as pairwise mutational fitness effects from the simulated sequence time series for short antigenic proteins. Our fitness inference approach may have potential future use for the design of immunization protocols by identifying intrinsically vulnerable immune target combinations on antigens that evolve under immune-driven selection. In the future, this approach may be applied to influenza and other novel viruses such as SARS-CoV-2, which evolves and, like influenza, might continue to escape the natural and vaccine-mediated immune pressures.

Comparative Population Genetics in the Human Gut Microbiome

Genetic variation in the human gut microbiome is responsible for conferring a number of crucial phenotypes like the ability to digest food and metabolize drugs. Yet, our understanding of how this variation arises and is maintained remains relatively poor. Thus, the microbiome remains a largely untapped resource, as the large number of coexisting species in the microbiome presents a unique opportunity to compare and contrast evolutionary processes across species to identify universal trends and deviations. Here we outline features of the human gut microbiome that, while not unique in isolation, as an assemblage make it a system with unparalleled potential for comparative population genomics studies. We consciously take a broad view of comparative population genetics, emphasizing how sampling a large number of species allows researchers to identify universal evolutionary dynamics in addition to new genes, which can then be leveraged to identify exceptional species that deviate from general patterns. To highlight the potential power of comparative population genetics in the microbiome, we reanalyze patterns of purifying selection across ∼40 prevalent species in the human gut microbiome to identify intriguing trends which highlight functional categories in the microbiome that may be under more or less constraint.

A Two-Level, Intramutant Spectrum Haplotype Profile of Hepatitis C Virus Revealed by Self-Organized Maps

RNA viruses replicate as complex mutant spectra termed viral quasispecies. The frequency of each individual genome in a mutant spectrum depends on its rate of generation and its relative fitness in the replicating population ensemble. The advent of deep sequencing methodologies allows for the first-time quantification of haplotype abundances within mutant spectra. There is no information on the haplotype profile of the resident genomes and how the landscape evolves when a virus replicates in a controlled cell culture environment. Here, we report the construction of intramutant spectrum haplotype landscapes of three amplicons of the NS5A-NS5B coding region of hepatitis C virus (HCV). Two-dimensional (2D) neural networks were constructed for 44 related HCV populations derived from a common clonal ancestor that was passaged up to 210 times in human hepatoma Huh-7.5 cells in the absence of external selective pressures. The haplotype profiles consisted of an extended dense basal platform, from which a lower number of protruding higher peaks emerged. As HCV increased its adaptation to the cells, the number of haplotype peaks within each mutant spectrum expanded, and their distribution shifted in the 2D network. The results show that extensive HCV replication in a monotonous cell culture environment does not limit HCV exploration of sequence space through haplotype peak movements. The landscapes reflect dynamic variation in the intramutant spectrum haplotype profile and may serve as a reference to interpret the modifications produced by external selective pressures or to compare with the landscapes of mutant spectra in complex in vivo environments. IMPORTANCE The study provides for the first time the haplotype profile and its variation in the course of virus adaptation to a cell culture environment in the absence of external selective constraints. The deep sequencing-based self-organized maps document a two-layer haplotype distribution with an ample basal platform and a lower number of protruding peaks. The results suggest an inferred intramutant spectrum fitness landscape structure that offers potential benefits for virus resilience to mutational inputs.

Redundancy-selection trade-off in phenotype-structured populations

Realistic fitness landscapes generally display a redundancy-fitness trade-off: highly fit trait configurations are inevitably rare, while less fit trait configurations are expected to be more redundant. The resulting sub-optimal patterns in the fitness distribution are typically described by means of effective formulations, where redundancy provided by the presence of neutral contributions is modelled implicitly, e.g. with a bias of the mutation process. However, the extent to which effective formulations are compatible with explicitly redundant landscapes is yet to be understood, as well as the consequences of a potential miss-match. Here we investigate the effects of such trade-off on the evolution of phenotype-structured populations, characterised by continuous quantitative traits. We consider a typical replication-mutation dynamics, and we model redundancy by means of two dimensional landscapes displaying both selective and neutral traits. We show that asymmetries of the landscapes will generate neutral contributions to the marginalised fitness-level description, that cannot be described by effective formulations, nor disentangled by the full trait distribution. Rather, they appear as effective sources, whose magnitude depends on the geometry of the landscape. Our results highlight new important aspects on the nature of sub-optimality. We discuss practical implications for rapidly mutant populations such as pathogens and cancer cells, where the qualitative knowledge of their trait and fitness distributions can drive disease management and intervention policies.

Hidden role of mutations in the evolutionary process

Mutations not only alter allele frequencies in a genetic pool but may also determine the fate of an evolutionary process. Here we study which allele fixes in a one-step, one-way model including the wild type and two adaptive mutations. We study the effect of the four basic evolutionary mechanisms-genetic drift, natural selection, mutation, and gene flow-on mutant fixation and its kinetics. Determining which allele is more likely to fix is not simply a question of comparing fitnesses and mutation rates. For instance, if the allele of interest is less fit than the other, then not only must it have a greater mutation rate, but also its mutation rate must exceed a specific threshold for it to prevail. We find exact expressions for such conditions. Our conclusions are based on the mathematical description of two extreme but important regimes, as well as on simulations.

A scaling law of multilevel evolution: how the balance between within- and among-collective evolution is determined

Numerous living systems are hierarchically organised, whereby replicating components are grouped into reproducing collectives-e.g., organelles are grouped into cells, and cells are grouped into multicellular organisms. In such systems, evolution can operate at two levels: evolution among collectives, which tends to promote selfless cooperation among components within collectives (called altruism), and evolution within collectives, which tends to promote cheating among components within collectives. The balance between within- and among-collective evolution thus exerts profound impacts on the fitness of these systems. Here, we investigate how this balance depends on the size of a collective (denoted by N) and the mutation rate of components (m) through mathematical analyses and computer simulations of multiple population genetics models. We first confirm a previous result that increasing N or m accelerates within-collective evolution relative to among-collective evolution, thus promoting the evolution of cheating. Moreover, we show that when within- and among-collective evolution exactly balance each other out, the following scaling relation generally holds: Nmα is a constant, where scaling exponent α depends on multiple parameters, such as the strength of selection and whether altruism is a binary or quantitative trait. This relation indicates that although N and m have quantitatively distinct impacts on the balance between within- and among-collective evolution, their impacts become identical if m is scaled with a proper exponent. Our results thus provide a novel insight into conditions under which cheating or altruism evolves in hierarchically-organised replicating systems.

Genealogical structure changes as range expansions transition from pushed to pulled

Range expansions accelerate evolution through multiple mechanisms, including gene surfing and genetic drift. The inference and control of these evolutionary processes ultimately rely on the information contained in genealogical trees. Currently, there are two opposing views on how range expansions shape genealogies. In invasion biology, expansions are typically approximated by a series of population bottlenecks producing genealogies with only pairwise mergers between lineages-a process known as the Kingman coalescent. Conversely, traveling wave models predict a coalescent with multiple mergers, known as the Bolthausen-Sznitman coalescent. Here, we unify these two approaches and show that expansions can generate an entire spectrum of coalescent topologies. Specifically, we show that tree topology is controlled by growth dynamics at the front and exhibits large differences between pulled and pushed expansions. These differences are explained by the fluctuations in the total number of descendants left by the early founders. High growth cooperativity leads to a narrow distribution of reproductive values and the Kingman coalescent. Conversely, low growth cooperativity results in a broad distribution, whose exponent controls the merger sizes in the genealogies. These broad distribution and non-Kingman tree topologies emerge due to the fluctuations in the front shape and position and do not occur in quasi-deterministic simulations. Overall, our results show that range expansions provide a robust mechanism for generating different types of multiple mergers, which could be similar to those observed in populations with strong selection or high fecundity. Thus, caution should be exercised in making inferences about the origin of non-Kingman genealogies.

Adaptation in a heterogeneous environment I: persistence versus extinction

Understanding how a diversity of plants in agroecosystems affects the adaptation of pathogens is a key issue in agroecology. We analyze PDE systems describing the dynamics of adaptation of two phenotypically structured populations, under the effects of mutation, selection and migration in a two-patch environment, each patch being associated with a different phenotypic optimum. We consider two types of growth functions that depend on the n-dimensional phenotypic trait: either local and linear or nonlocal nonlinear. In both cases, we obtain existence and uniqueness results as well as a characterization of the large-time behaviour of the solution (persistence or extinction) based on the sign of a principal eigenvalue. We show that migration between the two environments decreases the chances of persistence, with in some cases a 'lethal migration threshold' above which persistence is not possible. Comparison with stochastic individual-based simulations shows that the PDE approach accurately captures this threshold. Our results illustrate the importance of cultivar mixtures for disease prevention and control.

Antigenic waves of virus-immune coevolution

The evolution of many microbes and pathogens, including circulating viruses such as seasonal influenza, is driven by immune pressure from the host population. In turn, the immune systems of infected populations get updated, chasing viruses even farther away. Quantitatively understanding how these dynamics result in observed patterns of rapid pathogen and immune adaptation is instrumental to epidemiological and evolutionary forecasting. Here we present a mathematical theory of coevolution between immune systems and viruses in a finite-dimensional antigenic space, which describes the cross-reactivity of viral strains and immune systems primed by previous infections. We show the emergence of an antigenic wave that is pushed forward and canalized by cross-reactivity. We obtain analytical results for shape, speed, and angular diffusion of the wave. In particular, we show that viral-immune coevolution generates an emergent timescale, the persistence time of the wave's direction in antigenic space, which can be much longer than the coalescence time of the viral population. We compare these dynamics to the observed antigenic turnover of influenza strains, and we discuss how the dimensionality of antigenic space impacts the predictability of the evolutionary dynamics. Our results provide a concrete and tractable framework to describe pathogen-host coevolution.

Uniqueness of RNA Coliphage Qβ Display System in Directed Evolutionary Biotechnology

Phage display technology involves the surface genetic engineering of phages to expose desirable proteins or peptides whose gene sequences are packaged within phage genomes, thereby rendering direct linkage between genotype with phenotype feasible. This has resulted in phage display systems becoming invaluable components of directed evolutionary biotechnology. The M13 is a DNA phage display system which dominates this technology and usually involves selected proteins or peptides being displayed through surface engineering of its minor coat proteins. The displayed protein or peptide’s functionality is often highly reduced due to harsh treatment of M13 variants. Recently, we developed a novel phage display system using the coliphage Qβ as a nano-biotechnology platform. The coliphage Qβ is an RNA phage belonging to the family of Leviviridae, a long investigated virus. Qβ phages exist as a quasispecies and possess features making them comparatively more suitable and unique for directed evolutionary biotechnology. As a quasispecies, Qβ benefits from the promiscuity of its RNA dependent RNA polymerase replicase, which lacks proofreading activity, and thereby permits rapid variant generation, mutation, and adaptation. The minor coat protein of Qβ is the readthrough protein, A₁. It shares the same initiation codon with the major coat protein and is produced each time the ribosome translates the UGA stop codon of the major coat protein with the of misincorporation of tryptophan. This misincorporation occurs at a low level (1/15). Per convention and definition, A₁ is the target for display technology, as this minor coat protein does not play a role in initiating the life cycle of Qβ phage like the pIII of M13. The maturation protein A₂ of Qβ initiates the life cycle by binding to the pilus of the F⁺ host bacteria. The extension of the A₁ protein with a foreign peptide probe recognizes and binds to the target freely, while the A₂ initiates the infection. This avoids any disturbance of the complex and the necessity for acidic elution and neutralization prior to infection. The combined use of both the A₁ and A₂ proteins of Qβ in this display system allows for novel bio-panning, in vitro maturation, and evolution. Additionally, methods for large library size construction have been improved with our directed evolutionary phage display system. This novel phage display technology allows 12 copies of a specific desired peptide to be displayed on the exterior surface of Qβ in uniform distribution at the corners of the phage icosahedron. Through the recently optimized subtractive bio-panning strategy, fusion probes containing up to 80 amino acids altogether with linkers, can be displayed for target selection. Thus, combined uniqueness of its genome, structure, and proteins make the Qβ phage a desirable suitable innovation applicable in affinity maturation and directed evolutionary biotechnology. The evolutionary adaptability of the Qβ phage display strategy is still in its infancy. However, it has the potential to evolve functional domains of the desirable proteins, glycoproteins, and lipoproteins, rendering them superior to their natural counterparts.

The evolutionary origin of the universal distribution of mutation fitness effect

An intriguing fact long defying explanation is the observation of a universal exponential distribution of beneficial mutations in fitness effect for different microorganisms. To explain this effect, we use a population model including mutation, directional selection, linkage, and genetic drift. The multiple-mutation regime of adaptation at large population sizes (traveling wave regime) is considered. We demonstrate analytically and by simulation that, regardless of the inherent distribution of mutation fitness effect across genomic sites, an exponential distribution of fitness effects emerges in the long term. This result follows from the exponential statistics of the frequency of the less-fit alleles, f, that we predict to evolve, in the long term, for both polymorphic and monomorphic sites. We map the logarithmic slope of the distribution onto the previously derived fixation probability and demonstrate that it increases linearly in time. Our results demonstrate a striking difference between the distribution of fitness effects observed experimentally for naturally occurring mutations, and the "inherent" distribution obtained in a directed-mutagenesis experiment, which can have any shape depending on the organism. Based on these results, we develop a new method to measure the fitness effect of mutations for each variable residue using DNA sequences sampled from adapting populations. This new method is not sensitive to linkage effects and does not require the one-site model assumptions.

Mutation bias can shape adaptation in large asexual populations experiencing clonal interference

The extended evolutionary synthesis invokes a role for development in shaping adaptive evolution, which in population genetics terms corresponds to mutation-biased adaptation. Critics have claimed that clonal interference makes mutation-biased adaptation rare. We consider the behaviour of two simultaneously adapting traits, one with larger mutation rate U, the other with larger selection coefficient s, using asexual travelling wave models. We find that adaptation is dominated by whichever trait has the faster rate of adaptation v in isolation, with the other trait subject to evolutionary stalling. Reviewing empirical claims for mutation-biased adaptation, we find that not all occur in the 'origin-fixation' regime of population genetics where v is only twice as sensitive to s as to U. In some cases, differences in U are at least ten to twelve times larger than differences in s, as needed to cause mutation-biased adaptation even in the 'multiple mutations' regime. Surprisingly, when U > s in the 'diffusive-mutation' regime, the required sensitivity ratio is also only two, despite pervasive clonal interference. Given two traits with identical v, the benefit of having higher s is surprisingly small, occurring largely when one trait is at the boundary between the origin-fixation and multiple mutations regimes.

Velocity fluctuations of stochastic reaction fronts propagating into an unstable state: Strongly pushed fronts

The empirical velocity of a reaction-diffusion front, propagating into an unstable state, fluctuates because of the shot noises of the reactions and diffusion. Under certain conditions these fluctuations can be described as a diffusion process in the reference frame moving with the average velocity of the front. Here we address pushed fronts, where the front velocity in the deterministic limit is affected by higher-order reactions and is therefore larger than the linear spread velocity. For a subclass of these fronts-strongly pushed fronts-the effective diffusion constant D_{f}∼1/N of the front can be calculated, in the leading order, via a perturbation theory in 1/N≪1, where N≫1 is the typical number of particles in the transition region. This perturbation theory, however, overestimates the contribution of a few fast particles in the leading edge of the front. We suggest a more consistent calculation by introducing a spatial integration cutoff at a distance beyond which the average number of particles is of order 1. This leads to a nonperturbative correction to D_{f} which even becomes dominant close to the transition point between the strongly and weakly pushed fronts. At the transition point we obtain a logarithmic correction to the 1/N scaling of D_{f}. We also uncover another, and quite surprising, effect of the fast particles in the leading edge of the front. Because of these particles, the position fluctuations of the front can be described as a diffusion process only on very long time intervals with a duration Δt≫τ_{N}, where τ_{N} scales as N. At intermediate times the position fluctuations of the front are anomalously large and nondiffusive. Our extensive Monte Carlo simulations of a particular reacting lattice gas model support these conclusions.

A percolation model of natural selection

A new approach has been proposed and developed: the selection of optimal variants in the evolutionary mutation flow is considered as an analogue of a percolation filter. Interaction of mutations in a series of generations and random processes of drift determine the collective behavior of nodes (individuals - carriers and converters of mutations) and bonds (mutations) in the space of percolation lattice. It is shown that the choice of the development trajectory at the population level depends on the spectrum of supporting and prohibiting mutations under the influence of conjugate deterministic and random factors. From the point of view of the fluctuation-bifurcation process, new concepts of the lower and upper thresholds of the percolation selection grid are defined in the hierarchical structure of speciation. The upper threshold determines the state of self-organized criticality, which, when overcome, leads to irreversible self-organization processes in the population caused by the accumulation of mutations.

Virus population dynamics examined with experimental model systems

Experimental evolution permits exploring the effect of controlled environmental variables in virus evolution. Several designs in cell culture and in vivo have established basic concepts that can assist in the interpretation of evolutionary events in the field. Important information has come from cytolytic and persistent infections in cell culture that have unveiled the power of virus-cell coevolution in virus and cell diversification. Equally informative are comparisons of the response of viral populations when subjected to different passage régimens. In particular, plaque-to-plaque transfers in cell culture have revealed unusual genotypes and phenotypes that populate minority layers of viral quasispecies. Some of these viruses display properties that contradict features established in virology textbooks. Several hypotheses and principles of population genetics have found experimental confirmation in experimental designs with viruses. The possibilities of using experimental evolution to understand virus behavior are still largely unexploited.

High-resolution lineage tracking reveals travelling wave of adaptation in laboratory yeast

In rapidly adapting asexual populations, including many microbial pathogens and viruses, numerous mutant lineages often compete for dominance within the population^1-5. These complex evolutionary dynamics determine the outcomes of adaptation, but have been difficult to observe directly. Previous studies have used whole-genome sequencing to follow molecular adaptation^6-10; however, these methods have limited resolution in microbial populations. Here we introduce a renewable barcoding system to observe evolutionary dynamics at high resolution in laboratory budding yeast. We find nested patterns of interference and hitchhiking even at low frequencies. These events are driven by the continuous appearance of new mutations that modify the fates of existing lineages before they reach substantial frequencies. We observe how the distribution of fitness within the population changes over time, and find a travelling wave of adaptation that has been predicted by theory^11-17. We show that clonal competition creates a dynamical 'rich-get-richer' effect: fitness advantages that are acquired early in evolution drive clonal expansions, which increase the chances of acquiring future mutations. However, less-fit lineages also routinely leapfrog over strains of higher fitness. Our results demonstrate that this combination of factors, which is not accounted for in existing models of evolutionary dynamics, is critical in determining the rate, predictability and molecular basis of adaptation.

Phylodynamic theory of persistence, extinction and speciation of rapidly adapting pathogens

Rapidly evolving pathogens like influenza viruses can persist by changing their antigenic properties fast enough to evade the adaptive immunity, yet they rarely split into diverging lineages. By mapping the multi-strain Susceptible-Infected-Recovered model onto the traveling wave model of adapting populations, we demonstrate that persistence of a rapidly evolving, Red-Queen-like state of the pathogen population requires long-ranged cross-immunity and sufficiently large population sizes. This state is unstable and the population goes extinct or 'speciates' into two pathogen strains with antigenic divergence beyond the range of cross-inhibition. However, in a certain range of evolutionary parameters, a single cross-inhibiting population can exist for times long compared to the time to the most recent common ancestor ([Formula: see text]) and gives rise to phylogenetic patterns typical of influenza virus. We demonstrate that the rate of speciation is related to fluctuations of [Formula: see text] and construct a 'phase diagram' identifying different phylodynamic regimes as a function of evolutionary parameters.

Coupling adaptive molecular evolution to phylodynamics using fitness-dependent birth-death models

Beneficial and deleterious mutations cause the fitness of lineages to vary across a phylogeny and thereby shape its branching structure. While standard phylogenetic models do not allow mutations to feedback and shape trees, birth-death models can account for this feedback by letting the fitness of lineages depend on their type. To date, these multi-type birth-death models have only been applied to cases where a lineage’s fitness is determined by a single character state. We extend these models to track sequence evolution at multiple sites. This approach remains computationally tractable by tracking the genotype and fitness of lineages probabilistically in an approximate manner. Although approximate, we show that we can accurately estimate the fitness of lineages and site-specific mutational fitness effects from phylogenies. We apply this approach to estimate the population-level fitness effects of mutations in Ebola and influenza virus, and compare our estimates with in vitro fitness measurements for these mutations.

Clonal interference can cause wavelet-like oscillations of multilocus linkage disequilibrium

Within-host adaptation of pathogens such as human immunodeficiency virus (HIV) often occurs at more than two loci. Multiple beneficial mutations may arise simultaneously on different genetic backgrounds and interfere, affecting each other's fixation trajectories. Here, we explore how these evolutionary dynamics are mirrored in multilocus linkage disequilibrium (MLD), a measure of multi-way associations between alleles. In the parameter regime corresponding to HIV, we show that deterministic early infection models induce MLD to oscillate over time in a wavelet-like fashion. We find that the frequency of these oscillations is proportional to the rate of adaptation. This signature is robust to drift, but can be eroded by high variation in fitness effects of beneficial mutations. Our findings suggest that MLD oscillations could be used as a signature of interference among multiple equally advantageous mutations and may aid the interpretation of MLD in data.

Survival of the simplest in microbial evolution

The evolution of microbial and viral organisms often generates clonal interference, a mode of competition between genetic clades within a population. Here we show how interference impacts systems biology by constraining genetic and phenotypic complexity. Our analysis uses biophysically grounded evolutionary models for molecular phenotypes, such as fold stability and enzymatic activity of genes. We find a generic mode of phenotypic interference that couples the function of individual genes and the population’s global evolutionary dynamics. Biological implications of phenotypic interference include rapid collateral system degradation in adaptation experiments and long-term selection against genome complexity: each additional gene carries a cost proportional to the total number of genes. Recombination above a threshold rate can eliminate this cost, which establishes a universal, biophysically grounded scenario for the evolution of sex. In a broader context, our analysis suggests that the systems biology of microbes is strongly intertwined with their mode of evolution.

In asexual populations selection at different genomic loci can interfere with each other. Here, using a biophysical model of molecular evolution the authors show that interference results in long-term degradation of molecular function, an effect that strongly depends on genome size.

Epistasis detectably alters correlations between genomic sites in a narrow parameter window

Different genomic sites evolve inter-dependently due to the combined action of epistasis, defined as a non-multiplicative contribution of alleles at different loci to genome fitness, and the physical linkage of different loci in genome. Both epistasis and linkage, partially compensated by recombination, cause correlations between allele frequencies at the loci (linkage disequilibrium, LD). The interaction and competition between epistasis and linkage are not fully understood, nor is their relative sensitivity to recombination. Modeling an adapting population in the presence of random mutation, natural selection, pairwise epistasis, and random genetic drift, we compare the contributions of epistasis and linkage. For this end, we use a panel of haplotype-based measures of LD and their various combinations calculated for epistatic and non-epistatic pairs separately. We compute the optimal percentages of detected and false positive pairs in a one-time sample of a population of moderate size. We demonstrate that true interacting pairs can be told apart in a sufficiently short genome within a narrow window of time and parameters. Outside of this parameter region, unless the population is extremely large, shared ancestry of individual sequences generates pervasive stochastic LD for non-interacting pairs masking true epistatic associations. In the presence of sufficiently strong recombination, linkage effects decrease faster than those of epistasis, and the detection of epistasis improves. We demonstrate that the epistasis component of locus association can be isolated, at a single time point, by averaging haplotype frequencies over multiple independent populations. These results demonstrate the existence of fundamental restrictions on the protocols for detecting true interactions in DNA sequence sets.

Circuit-Host Coupling Induces Multifaceted Behavioral Modulations of a Gene Switch

Quantitative modeling of gene circuits is fundamentally important to synthetic biology, as it offers the potential to transform circuit engineering from trial-and-error construction to rational design and, hence, facilitates the advance of the field. Currently, typical models regard gene circuits as isolated entities and focus only on the biochemical processes within the circuits. However, such a standard paradigm is getting challenged by increasing experimental evidence suggesting that circuits and their host are intimately connected, and their interactions can potentially impact circuit behaviors. Here we systematically examined the roles of circuit-host coupling in shaping circuit dynamics by using a self-activating gene switch as a model circuit. Through a combination of deterministic modeling, stochastic simulation, and Fokker-Planck equation formalism, we found that circuit-host coupling alters switch behaviors across multiple scales. At the single-cell level, it slows the switch dynamics in the high protein production regime and enlarges the difference between stable steady-state values. At the population level, it favors cells with low protein production through differential growth amplification. Together, the two-level coupling effects induce both quantitative and qualitative modulations of the switch, with the primary component of the effects determined by the circuit's architectural parameters. This study illustrates the complexity and importance of circuit-host coupling in modulating circuit behaviors, demonstrating the need for a new paradigm-integrated modeling of the circuit-host system-for quantitative understanding of engineered gene networks.

Cryptic selection forces and dynamic heritability in generalized phenotypic evolution

Individuals with different phenotypes can have widely-varying responses to natural selection, yet many classical approaches to evolutionary dynamics emphasize only how a population's average phenotype increases in fitness over time. However, recent experimental results have produced examples of populations that have multiple fitness peaks, or that experience frequency-dependence that affects the direction and strength of selection on certain individuals. Here, we extend classical fitness gradient formulations of natural selection in order to describe the dynamics of a phenotype distribution in terms of its moments-such as the mean, variance, and skewness. The number of governing equations in our model can be adjusted in order to capture different degrees of detail about the population. We compare our simplified model to direct Wright-Fisher simulations of evolution in several canonical fitness landscapes, and we find that our model provides a low-dimensional description of complex dynamics not typically explained by classical theory, such as cryptic selection forces due to selection on trait ranges, time-variation of the heritability, and nonlinear responses to stabilizing or disruptive selection due to asymmetric trait distributions. In addition to providing a framework for extending general understanding of common qualitative concepts in phenotypic evolution - such as fitness gradients, selection pressures, and heritability - our approach has practical importance for studying evolution in contexts in which genetic analysis is infeasible.

Effective models and the search for quantitative principles in microbial evolution

Microbes evolve rapidly. Yet they do so in idiosyncratic ways, which depend on the specific mutations that are beneficial or deleterious in a given situation. At the same time, some population-level patterns of adaptation are strikingly similar across different microbial systems, suggesting that there may also be simple, quantitative principles that unite these diverse scenarios. We review the search for simple principles in microbial evolution, ranging from the biophysical level to emergent evolutionary dynamics. A key theme has been the use of effective models, which coarse-grain over molecular and cellular details to obtain a simpler description in terms of a few effective parameters. Collectively, these theoretical approaches provide a set of quantitative principles that facilitate understanding, prediction, and potentially control of evolutionary phenomena, though formidable challenges remain due to the ecological complexity of natural populations.

Antigenic evolution of viruses in host populations

To escape immune recognition in previously infected hosts, viruses evolve genetically in immunologically important regions. The host’s immune system responds by generating new memory cells recognizing the mutated viral strains. Despite recent advances in data collection and analysis, it remains conceptually unclear how epidemiology, immune response, and evolutionary factors interact to produce the observed speed of evolution and the incidence of infection. Here we establish a general and simple relationship between long-term cross-immunity, genetic diversity, speed of evolution, and incidence. We develop an analytic method fusing the standard epidemiological susceptible-infected-recovered approach and the modern virus evolution theory. The model includes the factors of strain selection due to immune memory cells, random genetic drift, and clonal interference effects. We predict that the distribution of recovered individuals in memory serotypes creates a moving fitness landscape for the circulating strains which drives antigenic escape. The fitness slope (effective selection coefficient) is proportional to the reproductive number in the absence of immunity R₀ and inversely proportional to the cross-immunity distance a, defined as the genetic distance of a virus strain from a previously infecting strain conferring 50% decrease in infection probability. Analysis predicts that the evolution rate increases linearly with the fitness slope and logarithmically with the genomic mutation rate and the host population size. Fitting our analytic model to data obtained for influenza A H3N2 and H1N1, we predict the annual infection incidence within a previously estimated range, (4-7)%, and the antigenic mutation rate of U_b = (5 − 8) ⋅ 10⁻⁴ per transmission event per genome. Our prediction of the cross-immunity distance of a = (14 − 15) aminoacid substitutions agrees with independent data for equine influenza.

Spread of many RNA viruses in a population represents a competition between host immune responses and viral evolution. RNA viruses accumulate mutations in immunologically important regions to escape immune recognition in hosts previously exposed to infection, while the immune system responds by producing new memory cells. Despite recent advances in data collection and their analysis, it remains conceptually unclear how epidemiology, immune response, and evolutionary factors interact to produce the observed speed of evolution and its incidence. By combining the standard epidemiological approach with the modern theory of viral evolution, we predict a general relationship between long-term cross-immunity, antigenic diversity of virus, its evolution speed, infection incidence, and the time to the most recent common ancestor. We apply these theoretical findings to available data on influenza virus to determine two important parameters of its evolution and confirm the model. Current strategies of vaccination against influenza should take into account stochastic fluctuations in fitness effect of mutations predicted by the theory.

Immune response and within-host viral evolution: Immune response can accelerate evolution

The objectives of this paper are to explore the impact of immune response on within-host viral evolution towards higher Darwinian fitness and, in particular, to verify a hypothesis that immune response, which is insufficient to annihilate a viral infection, can accelerate this evolution. To address this issue, a model of within-host viral evolution with immune response is formulated. This model is an extension of a continuous phenotype space model of viral evolution that was earlier suggested by A. Korobeinikov and C. Dempsey, which incorporates strain-specific immune response with cross-immunity. The model is based upon Nowak-May and Wodarz models of within-host HIV dynamics and is mechanistic (based upon first principles); this allows straightforward interpretation of the model's parameters and simulation results, as well as its further developments. In order to make the simulation results and conclusions robust and reliable and to ensure that they do not depend on the particularities of an immune response model, four different mathematical models of cell-mediated immune response are considered with the proposed model. Simulations confirmed that immune response, when it is unable to eliminate viruses, accelerates viral evolution.

Front propagation and clustering in the stochastic nonlocal Fisher equation

In this work, we study the problem of front propagation and pattern formation in the stochastic nonlocal Fisher equation. We find a crossover between two regimes: a steadily propagating regime for not too large interaction range and a stochastic punctuated spreading regime for larger ranges. We show that the former regime is well described by the heuristic approximation of the system by a deterministic system where the linear growth term is cut off below some critical density. This deterministic system is seen not only to give the right front velocity, but also predicts the onset of clustering for interaction kernels which give rise to stable uniform states, such as the Gaussian kernel, for sufficiently large cutoff. Above the critical cutoff, distinct clusters emerge behind the front. These same features are present in the stochastic model for sufficiently small carrying capacity. In the latter, punctuated spreading, regime, the population is concentrated on clusters, as in the infinite range case, which divide and separate as a result of the stochastic noise. Due to the finite interaction range, if a fragment at the edge of the population separates sufficiently far, it stabilizes as a new cluster, and the processes begins anew. The deterministic cutoff model does not have this spreading for large interaction ranges, attesting to its purely stochastic origins. We show that this mode of spreading has an exponentially small mean spreading velocity, decaying with the range of the interaction kernel.

Fluctuations uncover a distinct class of traveling waves

Epidemics, flame propagation, and cardiac rhythms are classic examples of reaction-diffusion waves that describe a switch from one alternative state to another. Only two types of waves are known: pulled, driven by the leading edge, and pushed, driven by the bulk of the wave. Here, we report a distinct class of semipushed waves for which both the bulk and the leading edge contribute to the dynamics. These hybrid waves have the kinetics of pushed waves, but exhibit giant fluctuations similar to pulled waves. The transitions between pulled, semipushed, and fully pushed waves occur at universal ratios of the wave velocity to the Fisher velocity. We derive these results in the context of a species invading a new habitat by examining front diffusion, rate of diversity loss, and fluctuation-induced corrections to the expansion velocity. All three quantities decrease as a power law of the population density with the same exponent. We analytically calculate this exponent, taking into account the fluctuations in the shape of the wave front. For fully pushed waves, the exponent is -1, consistent with the central limit theorem. In semipushed waves, however, the fluctuations average out much more slowly, and the exponent approaches 0 toward the transition to pulled waves. As a result, a rapid loss of genetic diversity and large fluctuations in the position of the front occur, even for populations with cooperative growth and other forms of an Allee effect. The evolutionary outcome of spatial spreading in such populations could therefore be less predictable than previously thought.

A mathematical model of marine bacteriophage evolution

To explore how particularities of a host cell-virus system, and in particular host cell replication, affect viral evolution, in this paper we formulate a mathematical model of marine bacteriophage evolution. The intrinsic simplicity of real-life phage-bacteria systems, and in particular aquatic systems, for which the assumption of homogeneous mixing is well justified, allows for a reasonably simple model. The model constructed in this paper is based upon the Beretta-Kuang model of bacteria-phage interaction in an aquatic environment (Beretta & Kuang 1998 Math. Biosci.149, 57-76. (doi:10.1016/S0025-5564(97)10015-3)). Compared to the original Beretta-Kuang model, the model assumes the existence of a multitude of viral variants which correspond to continuously distributed phenotypes. It is noteworthy that the model is mechanistic (at least as far as the Beretta-Kuang model is mechanistic). Moreover, this model does not include any explicit law or mechanism of evolution; instead it is assumed, in agreement with the principles of Darwinian evolution, that evolution in this system can occur as a result of random mutations and natural selection. Simulations with a simplistic linear fitness landscape (which is chosen for the convenience of demonstration only and is not related to any real-life system) show that a pulse-type travelling wave moving towards increasing Darwinian fitness appears in the phenotype space. This implies that the overall fitness of a viral quasi-species steadily increases with time. That is, the simulations demonstrate that for an uneven fitness landscape random mutations combined with a mechanism of natural selection (for this particular system this is given by the conspecific competition for the resource) lead to the Darwinian evolution. It is noteworthy that in this system the speed of propagation of this wave (and hence the rate of evolution) is not constant but varies, depending on the current viral fitness and the abundance of susceptible bacteria. A specific feature of the original Beretta-Kuang model is that this model exhibits a supercritical Hopf bifurcation, leading to the loss of stability and the rise of self-sustained oscillations in the system. This phenomenon corresponds to the paradox of enrichment in the system. It is remarkable that under the conditions that ensure the bifurcation in the Beretta-Kuang model, the viral evolution model formulated in this paper also exhibits a rise in self-sustained oscillations of the abundance of all interacting populations. The propagation of the travelling wave, however, remains stable under these conditions. The only visible impact of the oscillations on viral evolution is a lower speed of the evolution.

Beneficial mutation-selection dynamics in finite asexual populations: a free boundary approach

Using a free boundary approach based on an analogy with ice melting models, we propose a deterministic PDE framework to describe the dynamics of fitness distributions in the presence of beneficial mutations with non-epistatic effects on fitness. Contrarily to most approaches based on deterministic models, our framework does not rely on an infinite population size assumption, and successfully captures the transient as well as the long time dynamics of fitness distributions. In particular, consistently with stochastic individual-based approaches or stochastic PDE approaches, it leads to a constant asymptotic rate of adaptation at large times, that most deterministic approaches failed to describe. We derive analytic formulas for the asymptotic rate of adaptation and the full asymptotic distribution of fitness. These formulas depend explicitly on the population size, and are shown to be accurate for a wide range of population sizes and mutation rates, compared to individual-based simulations. Although we were not able to derive an analytic description for the transient dynamics, numerical computations lead to accurate predictions and are computationally efficient compared to stochastic simulations. These computations show that the fitness distribution converges towards a travelling wave with constant speed, and whose profile can be computed analytically.

Rapid adaptation in large populations with very rare sex: Scalings and spontaneous oscillations

Genetic exchange in microbes and other facultative sexuals can be rare enough that evolution is almost entirely asexual and populations almost clonal. But the benefits of genetic exchange depend crucially on the diversity of genotypes in a population. How very rare recombination together with the accumulation of new mutations shapes the diversity of large populations and gives rise to faster adaptation is still poorly understood. This paper analyzes a particularly simple model: organisms with two asexual chromosomes that can reassort during rare matings that occur at a rate r. The speed of adaptation for large population sizes, N, is found to depend on the ratio ∼log(Nr)∕log(N). For larger populations, the r needed to yield the same speed decreases as a power of N. Remarkably, the population undergoes spontaneous oscillations alternating between phases when the fittest individuals are created by mutation and when they are created by reassortment, which - in contrast to conventional regimes - decreases the diversity. Between the two phases, the mean fitness jumps rapidly. The oscillatory dynamics and the strong fluctuations this induces have implications for the diversity and coalescent statistics. The results are potentially applicable to large microbial populations, especially viruses that have a small number of chromosomes. Some of the key features may be more broadly applicable for large populations with other types of rare genetic exchange.

A Population Dynamics Model for Clonal Diversity in a Germinal Center

Germinal centers (GCs) are micro-domains where B cells mature to develop high affinity antibodies. Inside a GC, B cells compete for antigen and T cell help, and the successful ones continue to evolve. New experimental results suggest that, under identical conditions, a wide spectrum of clonal diversity is observed in different GCs, and high affinity B cells are not always the ones selected. We use a birth, death and mutation model to study clonal competition in a GC over time. We find that, like all evolutionary processes, diversity loss is inherently stochastic. We study two selection mechanisms, birth-limited and death limited selection. While death limited selection maintains diversity and allows for slow clonal homogenization as affinity increases, birth limited selection results in more rapid takeover of successful clones. Finally, we qualitatively compare our model to experimental observations of clonal selection in mice.

Stronger selection can slow down evolution driven by recombination on a smooth fitness landscape

Stronger selection implies faster evolution—that is, the greater the force, the faster the change. This apparently self-evident proposition, however, is derived under the assumption that genetic variation within a population is primarily supplied by mutation (i.e. mutation-driven evolution). Here, we show that this proposition does not actually hold for recombination-driven evolution, i.e. evolution in which genetic variation is primarily created by recombination rather than mutation. By numerically investigating population genetics models of recombination, migration and selection, we demonstrate that stronger selection can slow down evolution on a perfectly smooth fitness landscape. Through simple analytical calculation, this apparently counter-intuitive result is shown to stem from two opposing effects of natural selection on the rate of evolution. On the one hand, natural selection tends to increase the rate of evolution by increasing the fixation probability of fitter genotypes. On the other hand, natural selection tends to decrease the rate of evolution by decreasing the chance of recombination between immigrants and resident individuals. As a consequence of these opposing effects, there is a finite selection pressure maximizing the rate of evolution. Hence, stronger selection can imply slower evolution if genetic variation is primarily supplied by recombination.