Forward from the crossroads of ecology and evolution

Jennifer K. Rowntree, David M. Shuker, Richard F. Preziosi

Abstract

Community genetics is a synthesis of community ecology and evolutionary biology. It examines how genetic variation within a species affects interactions among species to change ecological community structure and diversity. The use of community genetics approaches has greatly expanded in recent years and the evidence for ecological effects of genetic diversity is growing. The goal of current community genetics research is to determine the circumstances in which, and the mechanisms by which community genetic effects occur and is the focus of the papers in this special issue. We bring a new group of researchers into the community genetics fold. Using a mixture of empirical research, literature reviews and theoretical development, we introduce novel concepts and methods that we hope will enable us to develop community genetics into the future.

1. The birth and development of community genetics

Community genetics was first suggested in print in 1992 by Janis Antonovics as a new area of research in a volume on the ecology, evolution and genetics of plant resistance to herbivores and pathogens [1]. The idea was born out of the recognition that most ecological systems do not involve simple pairwise interactions among species, but multiple complex interactions within and among trophic levels, and that traditional coevolutionary models were too narrow to explain species interactions in this context. The purpose of community genetics was to examine the ‘evolutionary genetic processes that occur among interacting populations in communities’ and to remove the need for reciprocal effects among interacting species demanded by traditional coevolutionary studies.

Antonovics [1] suggested two alternative approaches for studying community genetic effects, one holistic and one reductionist. He illustrated two viewpoints that fit neatly into these different approaches: (i) community genetics enables a focus on the effects of within-species genetic variation on ecological communities, and (ii) community genetics allows community processes to be viewed as the result of interactions among genotypes of different species. The latter (reductionist) view stems out of the plant resistance literature and applies population and quantitative genetic experimental approaches to more complex species interactions. The former (holistic) approach is more akin to community ecology, but focused on the effects of genetic variation on natural communities.

An important milestone in the development of community genetics as an autonomous research area was the 2003 Ecology special feature edited by A. A. Agrawal [2]. This consisted of two main papers by Whitham et al. [3] and Neuhauser et al. [4] and a number of commentaries [512]. Whitham et al. [3] reviewed the evidence for the presence of an extended phenotype (sensu Dawkins [13]) in multiple ecological systems. They highlighted the role of genetic variation in important traits of dominant or keystone species in structuring natural communities, and argued that genetic variation was an important consideration in the conservation of viable natural communities. This and subsequent work of the group (e.g. [1420]) have gone a long way to demonstrate the potentially far-reaching effects of genetic variation in ecologically relevant traits on community structure and ecosystem function. It also filled the initial gap in a holistic approach to community genetics.

Neuhauser et al. [4] took a more applied perspective and looked at the impact of humans on natural communities. They maintained that community genetic approaches were most useful when three conditions applied: (i) communities were not at equilibrium, (ii) genetic variation was present in ecologically relevant traits, and (iii) strong selection was acting, and that situations where these conditions applied were increasing with anthropogenic activities. They used three examples to illustrate their point: the introduction of transgenic Bt crops and the evolution of resistance in invertebrate herbivores, habitat fragmentation and population persistence in the long-lived perennial Echinacea angustifolia, and the population genetic effect of the expansion of agricultural maize into North America along with its associated fungal pathogen Ustilago maydis. While the conditions set out therein define scenarios where the detection of community genetic effects are probable, such effects have since been demonstrated in communities that we might reasonably expect to be close to equilibrium (e.g. [21,22]). Neuhauser et al. [4] demonstrated that neither ecological nor evolutionary models alone could explain the response of communities under strong selection and highlighted areas of applied research where a community genetic approach could answer tangible questions with a large socio-economic impact. Sadly, this kind of research remains under-represented in the field.

A significant further publication was the Trends in Ecology and Evolution article by Johnson & Stinchcombe [23]. They defined community genetics within a broader spectrum of work that sought to combine ideas and methods from evolutionary biology and ecology, i.e. placed community genetics within a renewed synthesis of ecological and evolutionary theory. They made distinctions between the time-scales of the interactions under investigation and the predominance of either evolutionary or ecological questions. In their view, community genetics falls within the shorter time-scale of rapid microevolution and has a more ecological focus. Johnson & Stinchcombe [23] argued that it was no longer enough to look for statistically significant ecological responses to genetic variation or evolutionary processes, but that the relative importance of genetic effects compared with other factors that order ecological communities must now be established.

In 2008, Hughes et al. [24] published a comprehensive review of work on the ecological consequences of genetic variation. They asked three questions: (i) how does one study the effects of genetic diversity in ecology, (ii) what are the mechanisms by which genetic diversity affects ecological processes, and (iii) what is the evidence for ecological effects of genetic diversity? Substantial examples therein provided evidence for significant ecological effects of genetic diversity but, similarly to Johnson & Stinchcombe [23], the authors concluded that there was insufficient evidence on the relative importance of community genetic effects compared with environmental factors, and that little was known about the mechanisms through which these genetic effects manifest themselves in communities. Bailey et al. [25] began to address the first of these issues with a meta-analysis on the relative effect sizes of introgression between hybridizing species and within-species genetic diversity on individual, community and ecosystem-level traits. Broadly, they showed that the effect of genetic variation (within species and hybrid systems) decreases with increasing levels of organization, so that genetic variation has the greatest effect on traits of the individuals and lessening effects on ecosystem processes. However, in some systems, the effects of genetic variation can be large and percolate across the whole community [25], such that general patterns may not apply for all communities. While this work goes some way to address the question of how important community genetic effects are, the number of studies used for the meta-analysis was still relatively small, concentrated on North American ecosystems and contained a number of confounding variables (e.g. the types of environment sampled were closely linked to the groups of organisms studied).

It is clear that community genetics can mean different things to different people, particularly whether the primary focus (or initial motivation) is either the ecology or the genetics of a system. Over the past decade, a number of scientific sub-disciplines have emerged with the aim of unifying ecology and evolutionary biology (e.g. niche construction [26], geographical mosaic theory of coevolution (GMTC) [27] and eco-evolutionary dynamics (EED) [28]; see the final paper of this issue [29] for further discussion). Community genetics is, therefore, one formulation among many and it is reasonable to ask if the continuation of a separate field of research is strictly necessary. Community genetics focuses, on the one hand, on genetic variation and evolutionary change on an ecological time-scale as drivers of community structure and, on the other hand, on community context (both biotic and abiotic) as a selective force. This latter facet of community genetics has obvious overlaps with niche construction theory [26], which is predicated on the direct and indirect evolutionary effects that all organisms can have on each other. These niche construction effects are, in turn, caused by the constant changes that organisms make to components of the biotic and abiotic environments by their daily activities and ultimate demise. The GMTC concentrates on differential coevolutionary trajectories among species in multiple populations [27,30]. Although possibly the result of community genetic interactions, GMTC is defined in terms of coevolution, and does not easily encompass multiple species interactions where reciprocal coevolution need not occur. A recent special issue on EED described EED as the bridge between ecology and evolution [28]. The issue contained papers from what we would consider a community genetics perspective (e.g. [25,31]), but also included papers from a much broader remit (e.g. [32,33]).

All of these formulations have merits, not least that they may appeal to different constituencies: for example, community genetics may appeal to community ecologists interested in the structure and function of ecological communities, and niche construction is finding favour in the human behavioural sciences. Our wish here is not to advocate one approach over another, but rather to make two points. A plurality of approaches is useful if together they broaden how ecologists and evolutionary biologists view their own fields, i.e. together they hasten a conceptual unification of evolution and ecology. However, this process could have a cost if, for instance, different perspectives generate apparently ‘novel’ theory or predictions that already exist in other sub-disciplines. There are lessons to be learnt here from the field of social evolution theory, where different modelling paradigms addressing the same underlying (i.e. evolutionary) processes have not always appreciated the similarity between paradigms [34,35]. This resulted in ‘new’ mechanisms or approaches being shown to be special cases (or mathematical analogues) of existing theory (e.g. [36,37]).

From our perspective, community genetics sits within a well-defined area and offers a focused view on the role of genetic variation and evolutionary potential in ecological communities, without the constraints of coevolution or the adoption of new conceptual paradigms, as envisaged by niche construction theory. We return and hold to the original views of Antonovics [1], as we believe that these, set within the framework proposed by Johnson & Stinchcombe [23] provide clear and relevant definitions of the field. Ultimately, we see community genetics simply as the realization that the environment experienced by most individuals is largely a product of the expressed genomes of other species. In other words, not all individuals (or genotypes) of a species can be treated as equal, and this inequality potentially affects the outcome of both con- and hetero-specific interactions, as well as having cascading effects on the community. The determination of the circumstance in which, and the mechanisms by which, such effects occur is the goal of current community genetics research and the focus of the papers in this special issue.

2. Community genetics terminology

Regardless of the framework, clearly defined terminology is crucial for conceptual clarity, and this will always be an issue when bringing ecologists and geneticists together. For instance, when merging ecological and evolutionary genetics perspectives on indirect effects, it is important to realize that the term ‘indirect effect’ suggests different processes and different numbers of species in the two fields. In an ecological framework, Wootton [38] defines indirect ecological effects (IEEs) as interactions between two species that depend on a third species. In trophic systems, this implies the presence of a third species that mediates the interactions of the other two species (e.g. [39]). In an evolutionary genetic framework, Moore et al. [40] define indirect genetic effects (IGEs) as the genotype of one individual influencing the phenotype of an interacting individual of the same species (note that no third individual is involved). This idea has been expanded in a community genetics framework to interspecific IGEs (IIGEs) [41], where the genotype of one species influences the phenotype of an interacting species. Thus, indirect effects from an evolutionary genetic perspective might involve one (IGE) or two (IIGE) species, whereas indirect effects from an ecological perspective (IEE) are defined to involve three species.

The concept of the ‘cross-species’ genetic correlation also needs to be made clear. Many of the evolutionary outcomes of community genetic effects are the result of the build-up of genetic correlations between species, and Wolf et al. [42] defined these within a quantitative genetic framework. While the concept of a genetic correlation across entities that do not share a genome might seem strange, it is merely the equivalent of the statistical quantitative genetic description of a genetic correlation within a species. In other words, there is a genetic covariance between phenotypes, and these phenotypes can exist within the same species or across two or more species (it should be remembered that this genetic covariance need not be limited to just two interacting species). These across-species covariances arise owing to assortative associations between genotypes across the interacting species, that is, specific combinations of genotypes are found together more often than expected from a random association of the two species. Such associations could occur by two different mechanisms. The first possibility is that genotypes associate at random but there is subsequent selection for specific combinations of genotypes, causing them to increase in relative frequency as other less favoured combinations become less common or are eliminated from the community. This type of co-occurrence of genotypes could be described as passive assortative association (e.g. [43]). The second possibility is that genotypes of one species actively seek out, and differentially associate with, genotypes of the second species. This type of co-occurrence might be best described as active assortative association (e.g. [44,45]).

3. The objectives of this issue

Since its inception, community genetics research has burgeoned (see [24] for a comprehensive review). Until recently, most of the work has come from North America, but this special issue aims to add a number of different perspectives to the field. The idea for this issue began at a workshop held in 2008 sponsored by the Genetics Society, whose purpose was to bring European researchers together to discuss future research directions in community genetics. This was followed by another meeting in 2010 sponsored by the British Ecological Society and the papers presented here come from presentations and ideas from participants at these two meetings. Here, we bring a new group of researchers into the community genetics fold, and present research that (a) attempts to answer some outstanding questions in the field and (b) expands on recent ideas. We introduce novel concepts and methods that we hope will enable us to develop community genetics research further and present a view of the direction community genetics research should now take. The papers we present are a mixture of empirical research, literature reviews and theoretical development.

Zytynska et al. [21] begin by expanding the idea of the dominant species. They look at the effect of genetic variation within a tree species on associated epiphyte and arthropod communities (sensu Whitham [3,18]), but do so in a complex tropical forest community where no single tree species predominates. They demonstrate that more genetically similar individuals of the breadnut tree (Brosimum alicastrum) harbour more similar communities of epiphytes and arthropods. This work is an important illustration that it is possible to tease apart meaningful community genetic effects even in ‘complex’ field systems in the tropics. Iason et al. [46] continue along the theme of tree extended phenotype effects. They concentrate on the Scots pine (Pinus sylvestris), and examine the role of herbivory as a selection pressure influencing the composition of chemical defence compounds. This is a novel perspective, looking at the influence of the attendant community on the direction and rate of evolution in a focal species, and provides evidence of some of the functional mechanisms driving community genetic processes. The paper by Whitlock et al. [22] is embedded in the search to understand the forces driving community structure and diversity, and focuses on the role of intraspecific genetic diversity in species-rich grasslands. The first half of the paper reviews the progress of grassland community genetics, placing our current understanding of this system in the context of direct effects and IIGEs. Community genetics work on grasslands is unique in that genetic diversity has been manipulated in multiple species simultaneously in a semi-realistic system, and the authors are worthy advocates of this approach. The second half of the paper presents some preliminary results that show retention of genetic diversity to be positively associated with the abundance of an interacting species. These results provide a thought-provoking insight into the way grasslands may be structured by their inherent genetic and species diversities, and how this influences interactions within and among species.

Wolf et al. [47] explore the genetic basis of ecological interactions among individuals in a population at greater depth by looking at competition among genotypes of Arabidopsis thaliana. They examine the functional basis of how genetic variation in a focal individual maps onto traits of an interacting individual via IGEs, using a combination of path and factor analysis, or structural equation models (SEMs), and quantitative trait loci (QTL) mapping of recombinant inbred lines (RILs). They seek to determine whether previously observed pleiotropy between traits in focal and interacting plants is caused by measured or unmeasured traits of the focal plant. This study provides a first step towards determining the explicit mechanisms driving indirect effects, and the use of SEM provides a framework for deciphering which of the many possible mechanisms are important in sustaining complex systems of interacting species. Although Wolf et al. [47] use a single species model here, this approach could be easily adapted for the study of interspecific interactions among genotypes, such as those illustrated in the previous paper by Whitlock et al. [22]. Tétard-Jones et al. [48] also explore the genetic basis of IGEs, but in this case using a tritrophic system of rhizobacteria, barley (Hordeum vulgare) and aphids (Sitobion avenae). They examine how changes in the rhizobacterial environment can influence aphid fitness via the intermediate host plant by mapping aphid and barley traits, and the plasticity of those traits onto the barley genome across rhizobacterial environments. In doing so, the authors provide initial evidence for the idea that indirect effects and phenotypic plasticity may interact within a community genetic framework. Their results suggest that phenotypic plasticity may play a significant role in bi-directional species interactions.

Rowntree et al. [49] introduce a new system using a parasitic plant (Rhinanthus spp.) and a grass host plant (barley). Rhinanthus are keystone species that induce dramatic changes in grassland plant and arthropod community composition by altering the competitive balance between potential host species. The authors examine tolerance, virulence and resistance traits of host and parasite, and demonstrate that variation within both species changes the outcome of interactions between the two. They argue that within-species genetic variation in this system will probably affect total community species composition in grasslands and recommend the use of a community genetics framework to explore complex host–parasite interactions. This theme is taken up by Ferrari & Vavre [50], who review the recent literature on bacterial symbionts and advocate their use as ideal model systems for community genetics. They explore the use and suitability of primary (obligatory) and secondary (facultative) bacterial symbionts as model systems for community genetics. In many ways, symbionts can be seen as an extension of their host genotypes, and the authors argue that they provide an alternative mechanism by which adaptation can occur. By doing so, they turn the idea of the extended phenotype around, with the host becoming an extension of the community of symbionts, rather than the community an extension of the host. Particularly exciting is the suggestion that communities of symbionts would be ideal systems in which to explore multiple levels of selection. Goodnight [51] develops the theory of multi-level selection across species, suggesting mechanisms by which this might occur. He argues that migration and the way in which meta-communities are founded are fundamental in determining how, and at what level, selection acts. This develops the ideas of Wade [52] that communities can be founded from migrant or propagule pools, either by independent individuals or by mixed, predetermined groups of individuals. The symbiont communities of Ferrari & Vavre [50] are a good example of the latter. Goodnight [51] moves these ideas on from single to multi-species communities. He argues that selection at the level of the community occurs when the fitness of an individual is a function of community membership and that selection rarely occurs on traits of isolated species, but rather on traits expressed in the context of a community. One of the theoretical challenges for the future will be to model these interspecific interactions with a variety of modelling paradigms, and try to avoid the terminological and mathematical confusion that often hinders the development of emerging fields.

The comparative method is one of the key approaches to understanding the causes and consequences of natural selection [53]. Traditionally, such analyses have been undertaken at the species level, using phylogenetically informed statistical approaches to analyse correlations between traits and environmental variables across a group of species. However, it has become clear that comparisons across different populations of the same species also require the statistical rigour applied to cross-species comparisons. Different populations of the same species may be statistically non-independent for a number of reasons, including shared (phylo)genetic history, gene flow and (spatial) correlation of habitats. Stone et al. [54] consider the approaches available for community geneticists and other evolutionary biologists to unpick these confounding effects for among-population comparisons. For community geneticists keen to compare the ecological consequences of genetic variation within and among species, studying multiple communities in the wild will become increasingly crucial, and correct statistical analysis is a prerequisite for detecting effect sizes that may be small. Although one of the messages given by Stone et al. [54] is that a definitive approach is still elusive, their contribution opens the door for the development of methods for cross-population studies, both in community genetics and beyond.

In his paper, Moya-Laraño [55] moves community genetics forward from a focus on fairly simple systems, where relatively few species interact, to far more complex and realistic communities. He offers an alternative top-down approach by starting with community-level effects and working back to the specific interactions. This form of modelling begins to determine potential effects of adding genetic variation into complex interacting networks of species (food webs) and builds on the seminal work of Bell [56] on the evolution of trophic structure. He demonstrates that realistic levels of genetic variation in traits that affect predator–prey interactions, namely growth rate and date of emergence of offspring, can affect the structure, robustness and stability of the surrounding food web.

Gatehouse et al. [57] bring community genetics firmly into an applied context and return us to Neuhauser et al.'s [4] 2003 suggestion that anthropogenically disturbed systems are ideal for the study of community genetics. Gatehouse et al. [57] consider the community-wide effects of a very specific kind of community genetics experiment: genetically modified crop plants. In a spirited argument for the use of genetically modified organisms (GMOs), they describe the effects of human interventions that change the genotype(s) of a key species within a community, in this case crop species within an agricultural landscape. Using a series of case studies, they argue that GMOs in these settings do not create cascading community effects that damage other organisms, including beneficial organisms such as natural predators. While GMOs in agriculture may seem an extremely artificial situation, it is one with potentially important consequences (both economically and socially), and GMOs represent a significant and growing proportion of the productivity and biomass in many landscapes. As such, predictions from theoretical and empirical work in other, more natural community genetic systems may well inform these ongoing attempts to genetically engineer at the landscape scale. On the other hand, if the cascading effects of GMOs are as limited as Gatehouse et al. [57] contend, then we may also learn something more general about how genotypes of dominant species shape ecological and evolutionary effects across trophic levels and communities as a whole (albeit with the caveat that agricultural systems may be less complex than many other communities).

The final paper by Hersch-Green et al. [29] takes a step back to consider community genetics in terms of its history and possible future paths. Interestingly in light of the preceding paper, they consider whether community genetics effects will prove more unpredictable and context-specific than we may currently appreciate (in effect evoking the spirit of the GMTC). If this proves to be the case, then the ecological and evolutionary impacts of cross-species genetics will make the outcomes of environmental management, not to mention the release of greater numbers of GMOs, harder to predict. A very real challenge for community genetics is whether or not a predictive science can be forged at the crossroads of ecology and evolution, a challenge, it has been argued [58], that ecology itself has yet to meet.

4. Concluding thoughts

While the breadth of systems where community genetic effects have been documented is ever expanding and incorporates aquatic, terrestrial, natural, agricultural and model systems, the list is not exhaustive. Previous experience has shown that collection of large amounts of relatively simple data (such as heritability estimates) allows the development of a strong empirical base from which to test theory. For instance, although laboratory-based estimates of heritability are context-dependent and may not necessarily reflect the genetic architecture of traits in the wild (but see [59]), the richness of the data collected by the quantitative genetics community has provided great insight into broad patterns of genetic architecture [6062]. We would like to see a similar resource develop over the next decade for community genetic parameters which would similarly allow us to predict if, when and where interspecific genetic interactions are important and whether ecological context influences evolutionary trajectories of interacting species. Such a database of effects would allow us to answer the key question ‘how important are community genetic effects in the face of other common ecological causes of variation?’ While we agree whole heartedly with the sentiment to move the field on and establish the importance of evolutionary processes in ecological interactions and vice versa (as advocated by Hersch-Green et al. [29]), determining the presence of community genetic effects in multiple and varied systems will remain an important activity if the field is to become more predictive. In his initial essay, Antonovics [1] realized that a modern ecological laboratory required a complimentary suite of molecular and field skills. With the development of modern high-throughput sequencing techniques and accompanying bioinformatics expertise, we now have the technology available to carry out the integrative community genetic research as Antonovics [1] fully envisaged. The future is indeed bright for community genetics!

Acknowledgements

We thank all the authors for their hard work in contributing such interesting papers to the issue and gratefully acknowledge the support of the Genetics Society and the British Ecological Society in funding the workshop and meeting on Community Genetics.

Footnotes

References

View Abstract