A stochastic model of the dynamics of host-pathogen systems with mutation

In this paper a stochastic model of the dynamics of host-pathogen systems with mutation is constructed. In previous works deterministic models of host-pathogen systems with no mutation were considered. The evolution of the pathogen population in any generation of the host is formulated as a multidimensional birth and death process, while the evolution of genotypic frequencies in successive generations of the host is described by a solution of a nonlinear vector difference equation. A general solution of the differential equations of the multidimensional birth and death process is presented and expressions for the stationary distribution, whenever it exists, and the mean time to extinction, when absorbing states are present, are derived. Some answers to questions raised in the discussion of a previous paper (Mode, 1962) are also contained in this paper. The research reported in this paper was supported by the United States Atomic Energy Comission, Division of Biology and Medicine Project AT(45-1)-1729.     

Evolution of Pathogen Specialisation in a Host Metapopulation: Joint Effects of Host and Pathogen Dispersal

Metapopulation processes are important determinants of epidemiological and evolutionary dynamics in host-pathogen systems, and are therefore central to explaining observed patterns of disease or genetic diversity. In particular, the spatial scale of interactions between pathogens and their hosts is of primary importance because migration rates of one species can affect both spatial and temporal heterogeneity of selection on the other. In this study we developed a stochastic and discrete time simulation model to specifically examine the joint effects of host and pathogen dispersal on the evolution of pathogen specialisation in a spatially explicit metapopulation. We consider a plant-pathogen system in which the host metapopulation is composed of two plant genotypes. The pathogen is dispersed by air-borne spores on the host metapopulation. The pathogen population is characterised by a single life-history trait under selection, the infection efficacy. We found that restricted host dispersal can lead to high amount of pathogen diversity and that the extent of pathogen specialisation varied according to the spatial scale of host-pathogen dispersal. We also discuss the role of population asynchrony in determining pathogen evolutionary outcomes.     

Density-Dependent Growth and Life History Evolution of Polycyclic Leaf Pathogens: A Continuous Time Growth Model

A density-dependent growth model of a polycyclic leaf pathogen is analysed. Viability and fecundity of the pathogen are regulated by the current population density. Spores are produced continuously from a sporulating infection until death of the infection, and consequently, all age classes of infections are present at a certain point in time. The leaf area of the host varies with time. Evolution of life history strategies are studied by letting different pathogen genotypes compete with each other. Evolution of life history strategies and evolution of impact of disease are discussed in relation to the ecology of the host-pathogen system. The model is exemplified by Erysiphe graminis f. sp. hordei growing on Hordeum vulgare.     

Identification of an ancestral resistance gene cluster involved in the coevolution process between Phaseolus vulgaris and its fungal pathogen Colletotrichum lindemuthianum.

The recent cloning of plant resistance (R) genes and the sequencing of resistance gene clusters have shed light on the molecular evolution of R genes. However, up to now, no attempt has been made to correlate this molecular evolution with the host-pathogen coevolution process at the population level. Cross-inoculations were carried out between 26 strains of the fungal pathogen Colletotrichum lindemuthianum and 48 Phaseolus vulgaris plants collected in the three centers of diversity of the host species. A high level of diversity for resistance against the pathogen was revealed. Most of the resistance specificities were overcome in sympatric situations, indicating an adaptation of the pathogen to the local host. In contrast, plants were generally resistant to allopatric strains, suggesting that R genes that were efficient against exotic strains but had been overcome locally were maintained in the plant genome. These results indicated that coevolution processes between the two protagonists led to a differentiation for resistance in the three centers of diversity of the host. To improve our understanding of the molecular evolution of these different specificities, a recombinant inbred (RI) population derived from two representative genotypes of the Andean (JaloEEP558) and Mesoamerican (BAT93) gene pools was used to map anthracnose specificities. A gene cluster comprising both Andean (Co-y; Co-z) and Mesoamerican (Co-9) host resistance specificities was identified, suggesting that this locus existed prior to the separation of the two major gene pools of P. vulgaris. Molecular analysis revealed a high level of complexity at this locus. It harbors 11 restriction fragment length polymorphisms when R gene analog (RGA) clones are used. The relationship between the coevolution process and diversification of resistance specificities at resistance gene clusters is discussed.     

Concerted evolution in the repeats of an immunomodulating cell surface protein, SOWgp, of the human pathogenic fungi Coccidioides immitis and C. posadasii

Genome dynamics that allow pathogens to escape host immune responses are fundamental to our understanding of host-pathogen interactions. Here we present the first population-based study of the process of concerted evolution in the repetitive domain of a protein-coding gene. This gene, SOWgp, encodes the immunodominant protein in the parasitic phase of the human pathogenic fungi Coccidioides immitis and C. posadasii. We sequenced the entire gene from strains representing the geographic ranges of the two Coccidioides species. By using phylogenetic and genetic distance analyses we discovered that the repetitive part of SOWgp evolves by concerted evolution, predominantly by the mechanism of unequal crossing over. We implemented a mathematical model originally developed for multigene families to estimate the rate of homogenization and recombination of the repetitive array, and the results indicate that the pattern of concerted evolution is a result of homogenization of repeat units proceeding at a rate close to the nucleotide point mutation rate. The release of the SOWgp molecules by the pathogen during proliferation may mislead the host: we speculate that the pathogen benefits from concerted evolution of repeated domains in SOWgp by an enhanced ability to misdirect the host's immune system.     

Differential trends in the codon usage patterns in HIV-1 genes

Host-pathogen interactions underlie one of the most complex evolutionary phenomena resulting in continual adaptive genetic changes, where pathogens exploit the host's molecular resources for growth and survival, while hosts try to eliminate the pathogen. Deciphering the molecular basis of host-pathogen interactions is useful in understanding the factors governing pathogen evolution and disease propagation. In host-pathogen context, a balance between mutation, selection, and genetic drift is known to maintain codon bias in both organisms. Studies revealing determinants of the bias and its dynamics are central to the understanding of host-pathogen evolution. We considered the Human Immunodeficiency Virus (HIV) type 1 and its human host to search for evolutionary signatures in the viral genome. Positive selection is known to dominate intra-host evolution of HIV-1, whereas high genetic variability underlies the belief that neutral processes drive inter-host differences. In this study, we analyze the codon usage patterns of HIV-1 genomes across all subtypes and clades sequenced over a period of 23 years. We show presence of unique temporal correlations in the codon bias of three HIV-1 genes illustrating differential adaptation of the HIV-1 genes towards the host preferred codons. Our results point towards gene-specific translational selection to be an important force driving the evolution of HIV-1 at the population level.     

Hybrid maize breeding with doubled haploids: I. One-stage versus two-stage selection for testcross performance

Optimum allocation of resources is of fundamental importance for the efficiency of breeding programs. The objectives of our study were to (1) determine the optimum allocation for the number of lines and test locations in hybrid maize breeding with doubled haploids (DHs) regarding two optimization criteria, the selection gain ΔG _k and the probability P _k of identifying superior genotypes, (2) compare both optimization criteria including their standard deviations (SDs), and (3) investigate the influence of production costs of DHs on the optimum allocation. For different budgets, number of finally selected lines, ratios of variance components, and production costs of DHs, the optimum allocation of test resources under one- and two-stage selection for testcross performance with a given tester was determined by using Monte Carlo simulations. In one-stage selection, lines are tested in field trials in a single year. In two-stage selection, optimum allocation of resources involves evaluation of (1) a large number of lines in a small number of test locations in the first year and (2) a small number of the selected superior lines in a large number of test locations in the second year, thereby maximizing both optimization criteria. Furthermore, to have a realistic chance of identifying a superior genotype, the probability P _k of identifying superior genotypes should be greater than 75%. For budgets between 200 and 5,000 field plot equivalents, P _k > 75% was reached only for genotypes belonging to the best 5% of the population. As the optimum allocation for P _k (5%) was similar to that for ΔG _k , the choice of the optimization criterion was not crucial. The production costs of DHs had only a minor effect on the optimum number of locations and on values of the optimization criteria. C. Friedrich H. Longin and H. Friedrich Utz contributed equally to this work.     

Reversion of a resistance‐breaking mutation shows reversion costs and high virus diversity at necrotic local lesions

An instance of host range evolution relevant to plant virus disease control is resistance breaking. Resistance breaking can be hindered by across‐host fitness trade‐offs generated by negative effects of resistance‐breaking mutations on the virus fitness in susceptible hosts. Different mutations in pepper mild mottle virus (PMMoV) coat protein result in the breaking in pepper plants of the resistance determined by the L ³ resistance allele. Of these, mutation M138N is widespread in PMMoV populations, despite associated fitness penalties in within‐host multiplication and survival. The stability of mutation M138N was analysed by serial passaging in L ³ resistant plants. Appearance on passaging of necrotic local lesions (NLL), indicating an effective L ³ resistance, showed reversion to nonresistance‐breaking phenotypes was common. Most revertant genotypes had the mutation N138K, which affects the properties of the virus particle, introducing a penalty of reversion. Hence, the costs of reversion may determine the evolution of resistance‐breaking in addition to resistance‐breaking costs. The genetic diversity of the virus population in NLL was much higher than in systemically infected tissues, and included mutations reported to break L ³ resistance other than M138N. Infectivity assays on pepper genotypes with different L alleles showed high phenotypic diversity in respect to L alleles in NLL, including phenotypes not reported in nature. Thus, high diversity at NLL may potentiate the appearance of genotypes that enable the colonization of new host genotypes or species. Collectively, the results of this study contribute to better understanding the evolutionary dynamics of resistance breaking and host‐range expansions.     

Reversion of a resistance-breaking mutation shows reversion costs and high virus diversity at necrotic local lesions

An instance of host range evolution relevant to plant virus disease control is resistance breaking. Resistance breaking can be hindered by across-host fitness trade-offs generated by negative effects of resistance-breaking mutations on the virus fitness in susceptible hosts. Different mutations in pepper mild mottle virus (PMMoV) coat protein result in the breaking in pepper plants of the resistance determined by the L³ resistance allele. Of these, mutation M138N is widespread in PMMoV populations, despite associated fitness penalties in within-host multiplication and survival. The stability of mutation M138N was analysed by serial passaging in L³ resistant plants. Appearance on passaging of necrotic local lesions (NLL), indicating an effective L³ resistance, showed reversion to nonresistance-breaking phenotypes was common. Most revertant genotypes had the mutation N138K, which affects the properties of the virus particle, introducing a penalty of reversion. Hence, the costs of reversion may determine the evolution of resistance-breaking in addition to resistance-breaking costs. The genetic diversity of the virus population in NLL was much higher than in systemically infected tissues, and included mutations reported to break L³ resistance other than M138N. Infectivity assays on pepper genotypes with different L alleles showed high phenotypic diversity in respect to L alleles in NLL, including phenotypes not reported in nature. Thus, high diversity at NLL may potentiate the appearance of genotypes that enable the colonization of new host genotypes or species. Collectively, the results of this study contribute to better understanding the evolutionary dynamics of resistance breaking and host-range expansions.     

Superinfections can induce evolutionarily stable coexistence of pathogens

Parasites reproduce and are subject to natural selection at several different, but intertwined, levels. In the recent paper, Gilchrist and Coombs (Theor. Popul. Biol. 69:145–153, 2006) relate the between-host transmission in the context of an SI model to the dynamics within a host. They demonstrate that within-host selection may lead to an outcome that differs from the outcome of selection at the host population level. In this paper we combine the two levels of reproduction by considering the possibility of superinfection and study the evolution of the pathogen’s within-host reproduction rate p. We introduce a superinfection function φ = φ(p,q), giving the probability with which pathogens with trait q, upon transmission to a host that is already infected by pathogens with trait p, “take over” the host. We consider three cases according to whether the function q → φ(p,q) (i) has a discontinuity, (ii) is continuous, but not differentiable, or (iii) is differentiable in q = p. We find that in case (i) the within-host selection dominates in the sense that the outcome of evolution at the host population level coincides with the outcome of evolution in a single infected host. In case (iii), it is the transmission to susceptible hosts that dominates the evolution to the extent that the singular strategies are the same as when the possibility of superinfections is ignored. In the biologically most relevant case (ii), both forms of reproduction contribute to the value of a singular trait. We show that when φ is derived from a branching process variant of the submodel for the within-host interaction of pathogens and target cells, the superinfection functions fall under case (ii). We furthermore demonstrate that the superinfection model allows for steady coexistence of pathogen traits at the host population level, both on the ecological, as well as on the evolutionary time scale.      

Host plant as an organizer of microbial evolution in the beneficial symbioses

Evolution of beneficial plant–microbe symbioses is presented as a result of selective processes induced by hosts in the associated microbial populations. These processes ensure a success of “genuine mutualists” (which benefit the host, often at the expense of their own fitness) in competition with “symbiotic cheaters” (which consume the resources provided by host without expressing the beneficial traits). Using a mathematical model describing the cyclic microevolution of rhizobia–legume symbiosis, we suggest that the selective pressures in favor of N₂-fixing (Fix⁺) strains operate within the in planta bacterial population due to preferential allocation of C resources into Fix⁺ nodules (positive partners’ feedbacks). Under the clonal infection of nodules, Fix⁺ strains (“genuine mutualists”) are supported by the group (inter-deme, kin) selection while under the mixed infections, this selection is ineffective since the Fix⁺ strains are over-competed by Fix⁻ ones (“symbiotic cheaters”) in the nodular habitats. Nevertheless, under mixed infections, Fix⁺ strains may be supported due to the coevolutionary responses form plant population which induce the mutualism-specific types of natural (group, individual) selection including the frequency dependent selection implemented in rhizobia population during the competition for host infection. Using the model of multi-strain bacterial competition for inoculation of symbiotic (rhizospheric, nodular) habitats, we demonstrate that the individual selection in favor of host-specific mutualist genotypes is more intensive than in favor of non-host-specific genotypes correlating the experimental data on the coordinated increases of symbiotic efficiency and specificity in the rhizobia–legume coevolution. However, an overall efficiency of symbiotic system is maximal when the non-host-specific mutualists are present in rhizobia population, and selection in favor of these mutualists operating at the whole population level is of key importance for improving the symbiosis. Construction of the agronomically valuable plant–microbe systems should provide the optimization of host-specific versus non-host-specific mutualists’ composition in legume inoculants combined with the clonal penetration of these mutualists into the nodules.     

Microevolution in an electronic microcosm

The evolution of microbial populations in simple environments such as chemostats is still not fully understood. The classical interpretation of adaptation involves a process of successive substitution whereby a new dominant genotype arises by mutation from the genotype previously dominant and spreads more or less rapidly through the population until it is nearly fixed. The population is, thus, nearly uniform most of the time. Some observations suggest that the process may be more complicated, but it remains formidably difficult to assemble the phylogeny of an evolving culture in sufficient detail to be sure. We report experiments with an electronic microcosm inhabited by self-replicating computer programs whose phylogeny can be rendered completely transparent. The physiology of these programs is different in many respects from that of organic creatures, but their population biology has many features in common, including a very extensive, if not unbounded, range of variation. Experimental populations evolved through point mutations (many of which were quasi-neutral when they were viable) and through rearrangements that led to a change in genome size and often had large effects on fitness. As a general rule, smaller genomes execute fewer instructions in order to replicate, the rate of replication increases as the number of instructions executed declines, and the rate of replication in pure culture is a good predictor of success in mixture. When cultured with CPU (central processing unit) time as the sole limiting resource, smaller genomes, therefore, evolve as a correlated response to natural selection for faster replication. The genetic basis of adaptation was highly contingent and always differed in replicate experiments. The pattern of evolution depends on mutation rate. At low mutation rates of 0.01 per genome per generation or less, we observed classic periodic selection, with each dominant genotype descending from the previous dominant and rising to a frequency of 0.8 or more. At higher mutation rates of about 0.1 per genome per generation, the most abundant genotypes rarely exceeded a frequency of about 0.4, and rare genotypes present in a few copies comprised a large part of the population. New dominant genotypes did not usually descend directly from previous dominants but, instead, from one of the many rare or moderately abundant genotypes. We suggest that the conventional chemostat paradigm may hold only as a special case at very low mutation rates and that the dynamics and diversity of evolving populations, even in the simplest conditions, may be more complex than is usually recognized. Artificial genetic autoadaptive systems are likely to be useful in constructing theory for situations that lie beyond the boundary of conventional population genetics.     

MHC polymorphism under host-pathogen coevolution

The genes encoding major histocompatibility (MHC) molecules are among the most polymorphic genes known for vertebrates. Since MHC molecules play an important role in the induction of immune responses, the evolution of MHC polymorphism is often explained in terms of increased protection of hosts against pathogens. Two selective pressures that are thought to be involved are (1) selection favoring MHC heterozygous hosts, and (2) selection for rare MHC alleles by host-pathogen coevolution. We have developed a computer simulation of coevolving hosts and pathogens to study the relative impact of these two mechanisms on the evolution of MHC polymorphism. We found that heterozygote advantage per se is insufficient to explain the high degree of polymorphism at the MHC, even in very large host populations. Host-pathogen coevolution, on the other hand, can easily account for realistic polymorphisms of more than 50 alleles per MHC locus. Since evolving pathogens mainly evade presentation by the most common MHC alleles in the host population, they provide a selective pressure for a large variety of rare MHC alleles. Provided that the host population is sufficiently large, a large set of MHC alleles can persist over many host generations under host-pathogen coevolution, despite the fact that allele frequencies continuously change.Electronic Supplementary Material Supplementary material is available in the online version of this article at     

RARER NEED NOT BE BETTER IF COMMONER IS WORSE: FREQUENCY-DEPENDENT SELECTION FOR DEVELOPMENTAL TIME AT THE ALCOHOL DEHYDROGENASE LOCUS OF THE OLIVE FRUIT FLY,BACTROCERA OLEAE

Whereas the importance of frequency-dependent selection in life-history traits, behavioral characters and source allocation patterns is widely accepted, its role in governing biochemical and molecular polymorphisms remains poorly understood. Here we demonstrate a case of allozyme frequency-dependent selection. When olive fruit flies (Bactrocera oleae) are reared on an artificial larval medium, an allele at the alcohol dehydrogenase locus that is present in very low frequency in natural populations increases to about one-third in less than five generations. We show here that the time from the hatching of the egg to the eclosion of the adult is affected by the genotype composition of the larval population that grows in the same cup of food. Cultures consisting of one genotype only have the longest developmental time, and two-allele cultures in which the two homozygotes and the heterozygote occur in a 1:1:2 ratio show the shortest developmental time. Cultures with intermediate genotypic compositions show intermediate levels of developmental time. The results can be explained by assuming that the developmental time of a genotype depends on the frequency array of all genotypes in the larval population and is not merely a function of its own frequency. It is even possible that the developmental time of a genotype becomes longer as the genotype becomes rarer, yet the genotype will be favored because the developmental times of the competing genotypes become even longer owing to the associated increase of their frequencies. Given that developmental time is inversely related to fitness, this generates a frequency-dependent selection, with developmental times changing progressively until the population arrives at an equilibrium. One optimum population composition that provides a satisfactory fit to allele frequency changes in our experimental populations is when the two alleles occur in equal frequencies and genotypes are in Hardy-Weinberg proportions. We argue that this type of selection is consistent with the role of alcohol dehydrogenase as a detoxifying enzyme in a medium that undergoes continuous chemical changes during its use by the feeding larvae.     

The effect of host heterogeneity and parasite intragenomic interactions on parasite population structure

Understanding the processes that shape the genetic structure of parasite populations and the functional consequences of different parasite genotypes is critical for our ability to predict how an infection can spread through a host population and for the design of effective vaccines to combat infection and disease. Here, we examine how the genetic structure of parasite populations responds to host genetic heterogeneity. We consider the well-characterized molecular specificity of major histocompatibility complex binding of antigenic peptides to derive deterministic and stochastic models. We use these models to ask, firstly, what conditions favour the evolution of generalist parasite genotypes versus specialist parasite genotypes? Secondly, can parasite genotypes coexist in a population? We find that intragenomic interactions between parasite loci encoding antigenic peptides are pivotal in determining the outcome of evolution. Where parasite loci interact synergistically (i.e. the recognition of additional antigenic peptides has a disproportionately large effect on parasite fitness), generalist parasite genotypes are favoured. Where parasite loci act multiplicatively (have independent effects on fitness) or antagonistically (have diminishing effects on parasite fitness), specialist parasite genotypes are favoured. A key finding is that polymorphism is not stable and that, with respect to functionally important antigenic peptides, parasite populations are dominated by a single genotype.     

Interplay between humans and infective agents: a population genetic study

The genetic composition of present day human populations is determined largely by the interaction between the human host and infective agents. Therefore, theoretical analysis of the host-infective-agent system is required in order for us to be able to understand human evolution. Classical population genetics has been confined largely to analysing the interplay of various mechanisms, such as selection, mutation and drift, in one species at a time. Unfortunately, there have been few studies of such interactive systems. In the present investigation, these studies have been enlarged, with problems of human genetics in mind, by mathematical examination of a model in which a diploid host with three alleles interacts with a haploid infective agent with two alleles. The results are compared with those obtained from simpler models analysed in the past. The assumptions inherent in such “gene for gene” models and our results are discussed. An increase in the number of alleles appears to enhance the chances for the establishment of permanent genetic polymorphisms, improving genetic “elasticity” of a population for coping with changing challenges by various infective agents. Interaction between two haploid species leads to a loss of polymorphism in both of them and, hence, to a severe loss of evolutionary elasticity. The hypothesis that the evolution of diploidy might have been favoured by a selective advantage of diploid organisms interacting with environmental challenges, such as infective agents, is supported. Received: 6 October 1997 / Accepted: 26 November 1997     

A two-locus mutation—Selection model and some of its evolutionary implications

The deterministic properties of a two-locus model with mutation and selection have been investigated. The mutation process is unidirectional, and the model is so constructed that the genetic variation at one locus is selectively neutral in the absence of a mutant allele at the other locus. All genotypes with three or four mutant alleles are deleterious, while the double heterozygotes may have the same fitness as the standard genotype. If one of the mutant alleles becomes fixed in the population, then the other locus will show a regular one-locus mutation-selection balance. Such a boundary equilibrium may be unstable or stable in the full two-locus setting. In the symmetric case, which is analyzed in details, the population will either go to one of the two boundary equilibria, or to a fully polymorphic equilibrium at which both the mutant alleles are rare. The origin of reproductive separation between two populations via the fixation of complementary deleterious mutants at different loci, and the fixation of nonfunctional alleles at duplicated loci, are two biological processes which both can be studied with the present model. In the last part of the paper we show how the results from the deterministic analysis can be used to predict how different factors will influence the rates of evolution in these systems.     

Transmission rates and adaptive evolution of pathogens in sympatric heterogeneous plant populations

Diversification in agricultural cropping patterns is widely practised to delay the build-up of virulent races that can overcome host resistance in pathogen populations. This can lead to balanced polymorphism, but the long-term consequences of this strategy for the evolution of crop pathogen populations are still unclear. The widespread occurrence of sibling species and reproductively isolated sub-species among fungal and oomycete plant pathogens suggests that evolutionary divergence is common. This paper develops a mathematical model of host-pathogen interactions using a simple framework of two hosts to analyse the influences of sympatric host heterogeneity on the long-term evolutionary behaviour of plant pathogens. Using adaptive dynamics, which assumes that sequential mutations induce small changes in pathogen fitness, we show that evolutionary outcomes strongly depend on the shape of the trade-off curve between pathogen transmission on sympatric hosts. In particular, we determine the conditions under which the evolutionary branching of a monomorphic into a dimorphic population occurs, as well as the conditions that lead to the evolution of specialist (single host range) or generalist (multiple host range) pathogen populations.     

The acquisition of hypovirulence in host-pathogen systems with three trophic levels

A major focus of research on the dynamics of host-pathogen interactions has been the evolution of pathogen virulence, which is defined as the loss in host fitness due to infection. It is usually assumed that changes in pathogen virulence are the result of selection to increase pathogen fitness. However, in some cases, pathogens have acquired hypovirulence by themselves becoming infected with hyperparasites. For example, the chestnut blight fungus Cryphonectria parasitica has become hypovirulent in some areas by acquiring a double-stranded RNA hyperparasite that debilitates the pathogen, thereby reducing its virulence to the host. In this article, we develop and analyze a mathematical model of the dynamics of host-pathogen interactions with three trophic levels. The system may be dominated by either uninfected (virulent) or hyperparasitized (hypovirulent) pathogens, or by a mixture of the two. Hypovirulence may allow some recovery of the host population, but it can also harm the host population if the hyperparasite moves the transmission rate of the pathogen closer to its evolutionarily stable strategy. In the latter case, the hyperparasite is effectively a mutualist of the pathogen. Selection among hyperparasites will often minimize the deleterious effects, or maximize the beneficial effects, of the hyperparasite on the pathogen. Increasing the frequency of multiple infections of the same host individual promotes the acquisition of hypovirulence by increasing the opportunity for horizontal transmission of the hyperparasite. This effect opposes the usual theoretical expectation that multiple infections promote the evolution of more virulent pathogens via selection for rapid growth within hosts.