Abstracts:  NVTB - Meeting  2003

It seems obvious that natural selection would in general favor pathogens with maximized infectiousness. In the (simplest?) case where infection depends on direct contact between hosts (and where the pathogen does not influence the host contact behavior) evolution should thus maximize both the transmission rate and the infectious period.

In this presentation we will demonstrate how self-organization into complex spatio-temporal patterns can lead to selection towards an 'inherent tradeoff' between pathogen transmission rate and infectious period. In particular, even when pathogen transmission rate and infectious period are both allowed to increase, selection in the spatial model will favor intermediate values for both traits (instead of maximal values). We show that the 'inherent tradeoff function' coincides with the phase-transition between regular and irregular patterns, thus providing a strong example of so-called evolution towards self-organized criticality.

Interestingly, previously so-called 'explicit tradeoffs' have been postulated between transmission rate and infectious period to explain e.g. the persistence of virulent pathogens. The popular argument is that pathogens, in order to become more infectious, draw more heavily on the host's resources and thus increase their virulence (e.g. host mortality). Unfortunately, such explicit tradeoff functions are hard to quantify in existing pathogen-host systems, and they are likely to vary between species. We demonstrate that, independent of the exact shape of the explicit tradeoff function, the evolutionary attractor on this function typically is at the intersection of the function with the inherent tradeoff function. Thus, the evolutionary stable attractor in these systems is to a large extent independent of the exact shape of the explicit tradeoff function.

More complex situations can arise when the explicit and inherent tradeoff functions intersect more than once. In these cases alternative stable evolutionary attractors can exist, depending on initial conditions. Furthermore, if slightly larger mutations in phenotype are allowed (possibly caused by small genotypic mutations), we show that also 'evolutionary cycling' can occur, resembling the rock-scissors-paper game. These last results illustrate the limitations of the so-called mutual invisibility analysis, as in this case specific mutants are only fit to invade into the turbulent interface between competing resident types. We conclude that the pre-structuring power of spatial pattern formation in evolutionary dynamics can hardly be overestimated.

A multitude of models have been employed to investigate the feasibility of sympatric speciation. A factor that potentially plays a role in speciation but that is often overlooked in these models is the influence of learning processes. Learning could for instance play a role when a new niche is being colonised, because the learning of niche characteristics may cause niche-specific assortative mating. Several animal species learn about features of their environment which may be important in finding or attracting mates. Examples include brood-parasitic birds learning songs of the foster species, insects that have a learned preference to feed on certain host plants, or fish that imprint olfactorily on their local environment. We use a recurrence equation model to look into the effect of learning on the colonisation of new niches, which is a step necessary to achieve speciation. We show that the learned characteristics can either facilitate or prevent the colonisation of new niches. Whether the effect on colonisation is positive or negative depends on female fertility and on viability selection against alleles that are adapted to either of the niches.

The semelparous species are those who reproduce once in their life and then die. There are several year classes who can either coexist or outcompete each other. In this talk I'll try to explain why people are interested in these species. I'll try to find the most important and intriguing questions. I am not going to give all answers, but I'll try to summarize what people (including us) are have already understood and what should be understood yet. The symbol of this talk will be cicada with its occurrence period of 13 and 17 years.

Holling's disk equation is almost universally used to describe functional responses but has some serious shortcomings. 1. The underlying idea of handling time is little more than a rough substitute for satiation, the main factor which keeps a predator from eating each available prey. 2. If two or more prey species are involved, the disk equation is unable to deal with the predator's preferences, and the potentially important consequences this has for ecosystem behaviour. These weak points can be remedied by a satiation-based stochastic model not unlike the model by Metz and Van Batenburg (1984). The resulting functional response is governed by a delayed differential equation which, unfortunately, is rather intractable. This hampers application as a buiding block in larger ecosystem models. The problem can be overcome by approximating the model's true response by a (non-orthogonal) hyperbola: the H response. The approximation ia remarkably accurate, and has interesting properties. 1. Compared to the disk equation, the H response contains one extra parameter which defines a range of curves, with Type 1 and disk equation at both ends, and can be identified as a measure of prey size. The H response thus provides a unified treatment for the hitherto disjoint Type 1 and disk equation. 2. Prey preference comes naturally with the model, and requires no separate specification. In accordance with observed behaviour, and unlike the disk equation, the H response in mixed prey populations predicts a shift in diet composition if overall prey density is changed.

The notion of satiation as used in this model is robust; it does not require a definite physiological interpretation like stomach content or blood sugar value, and yet can be measured and expressed in terms of behavourial observables.

Phylogenetic trees based on gene repertoires are remarkably similar to the current consensus of life history. Yet it has been argued that the shared gene content is an unreliable phylogenetic signal because of horizontal gene transfer and independent loss. In an attempt to strengthen the phylogenetic signal, we iteratively remove orthologous groups with an inconsistent phylogenetic ditribution in the phylogeny. This gives small but significant improvements of the taxonomy. Being too strict in removing orthologous groups does however lead to a breakdown of the phylogenetic pattern. We conclude that the phylogenetic signal in the majority of orthologous groups is needed for a consistent phylogeny.

Resilience of an ecological system is defined by the velocity of the system as it returns to its equilibrium state after some perturbation. Since the system does not arrive exactly in the equilibrium in finite time, the definition is based on the time needed to decrease the distance to the equilibrium with some fraction. In this study it is found that for stochastic populations this arbitrariness disappears because the equilibrium point can be replaced by a small (confidence) domain containing the equilibrium. The size of this domain is a measure for the persistence of the system. This method is fully worked out for the stochastic logistic equation as well as for a prey-predator system.

Two recent books, one by Gould (1989), Wonderful Life, and the other by Conway Morris (1998), The Crucible of Creation, have revived an old controversy in palaeontology, that concerning the existence of phylogenetic trends. Gould held that phylogenetic development is contingent, driven by stochastic extinction, contrary to Conway Morris who debated this, proposing the existence of phylogenetic trends instead. According to the first vision, the 'tape of life' would never give the same results, whereas according to the second, it would.

Firstly, it is difficult to evaluate these viewpoints, the debaters possibly arguing from the viewpoints of different scales of variation. On one scale, a process may be treated most parsimoniously with stochastic models, whereas on another one, a deterministic model applies best. Up-scaling or down-scaling can be risky.

Secondly, one may hold that Gould's stochastic approach concerns the statistical H₀ hypothesis and Conway Morris' the biological one of H₁. In that case, one has to test the possibility of the existence of a biologically meaningful trend according to the usual statistical methodology.

Thirdly, as seen from the viewpoint of the functionality of adaptation, one could also hold that these two statistical alternatives should, in fact, be integrated. Thus, biological adaptations can be viewed as a means reducing the effects of stochastic decay at various levels of variation, safeguarding the persistence of the system concerned.

Ecological experiments often accumulate data by carrying out many replicate trials, each containing a limited numbers of observations, which are then pooled and analysed for a pattern. This approach disregards the possibility that the overall pattern is influenced by different experimental conditions during replicate trials. This paper is about how to deal with such factors in model selection. Three methods of model selection are introduced: likelihood-ratio testing, the AIC with or without small-sample correction and the BIC. Subsequently, we apply the AICc method to an example on size-dependent seed dispersal by scatterhoarding rodents. The example involves binary data on harvesting and subsequent storage of Carapa nuts by rodents in plots within years of different ambient seed abundance. The question is whether there is an optimum seed size for harvesting and storage. We fit five models, varying from no effect of seed mass to an optimum seed mass. We show that lumping the data produces the expected pattern, but gives a poor fit compared to analyses in which grouping levels are taken into account. Model fitting for each plot separately gives the best fit but no general pattern. Using the minimum Kullback-Leibler distance as a measure for the difference between models, we explore how far models must differ in order to be able to discriminate between them. We then show by simulation that the differences between the five tested models are too small to discriminate between them at the plot level at all. We recommend a combined approach in which the level of lumping the trials is chosen by the amount of variation explained in comparison to an analysis at the trial level. It is shown that combining data of different plots only leads to an increase in the probability of identifying the correct model with the AIC criterion if the distance of all less extended models to the simulated model is sufficiently large in each plot. Otherwise, an increase in the number of replicate plots might lead to a decrease in the power of the AIC.

Though most phytoplankton competition studies have so far focussed on well-mixed systems, many natural phytoplankton communities are not well mixed. In particular, pronounced phytoplankton depth profiles may develop if local growth rates and/or transport rates of phytoplankton overcome the homogenizing effects of turbulent diffusion. Would slow mixing of the water column lead to a different outcome of competition than expected on the basis of well-mixed laboratory experiments? Here, we use a reaction-advection-diffusion model to investigate the effects of turbulent mixing on competition for light between sinking and buoyant phytoplankton species. In particular, we focus on the question which mixing regimes will select for species with high sinking rates, which mixing regimes will favor neutrally buoyant species, and which mixing regimes will favor positively buoyant species. The theoretical predictions will be compared with the results of a field experiment in which we changed the turbulence structure of an entire lake by artificial mixing.

Affinity maturation during immune responses to T-dependent antigens occurs in germinal centers (GC). In GCs antigen specific B cells undergo rounds of somatic mutations that alter their affinity. High affinity mutants take over GCs very soon after they appear; the replacement rate is as high as 4 per day. To gain more insight into this selection process, we present a spatial model of GC reactions, where B cells compete for survival signals from follicular dendritic cells (FDC). Assuming that high affinity B cells have increased cellular adhesion to FDCs, we obtain an affinity based sorting of B cells on the FDC. This sorting results in a winner-takes-all behavior. By comparing our sorting model with ``affinity-proportional selection models'', we show that a winner-takes-all selection is required to account for the fast rates at which high affinity mutants take over GCs. Thus, this model environment is one of the first that simulates winner-takes-all behaviour observed in vivo. Moreover, visual inspection of model GC cells is helpfull in getting a better understanding of GC reactions.

We investigate whether coexistence of plants is possible given the condition that there is only one limiting resource, namely light. Light is in principle a homogeneous resource, which means that there is no heterogeneity in light supply in the horizontal plane. Within the vegetation however, the plants themselves determine how much light is present. With their leaf area the plants intercept light and create a vertical light gradient. Allocation pattern and competitive interactions create the shape of a plant. This shape, in particular the height and the leaf area, determines how much light a plant intercepts. Our main question is whether plants with different allocation patterns are able to coexist. We use a mechanistic model to simulate competition between plants. In our model, plants possess different investment patterns in height. As biomass investments are drawn from the available pool of carbohydrates, investments in height cannot be used for an investment in leaf area, and vice versa. In this way leaf area is affected by the investment pattern in height. First I discuss differences in plant fitness between species pairs within a season and explain whether plants should be similar or dissimilar in order to have a more or less equal fitness. Also I discuss possibilities of coexistence between these species pairs when frequency is included and show how very distinct patterns of coexistence emerge. Finally I discuss possibilities of stable coexistence between multiple species with help of game theory.

The diel vertical migration (DVM) of Daphnia is generally interpreted as predator avoidance behaviour of the zooplankters against predation by fish. Numerous experiments and field observations are available about various aspects of this behaviour. The aim of our work was to bring these results into synthesis for a comprehensive understanding of the phenomenon DVM. An individual based model of daphnid DVM was developed to investigate the costs and benefits of diel vertical migration. The model consisted of three submodels. (1) The submodel 'movements' simulated individual migration movements in dependence of relative light change, extinction coefficient and fish biomass. Under daylight conditions the individuals performed a random walk for food search. (2) For the submodel 'growth and reproduction' a multiple regression approach was used to estimate somatic growth and reproduction in dependence of ambient temperature and food supply. Finally, (3) the submodel 'mortality' calculated individual probability of dying per time step. As sources of mortality size selective feeding and ageing were applied. The model was able to reproduce the daily migration behavior and the spatio-temporal distribution of the population in good agreement with literature data. Vertical migrations were performed in roughly two hours during dust and dawn. Population growth rate of migrating populations were reduced mainly due to temperature reduction during daytime. The model results support the 'predator avoidance theory' and showed high adaptive benefits under severe fish predation. Moreover, adaptive benefits were still detectable under relatively low fish biomasses. These adaptive benefits were further influenced by the vertical structure of the habitat in terms of food and temperature. Hence, a vertically heterogeneous structured environment, e.g. a deep chlorophyll maximum, was able to modify the optimal migration schedule of daphnids.

The assumption that trade-offs exist is fundamental in evolutionary theory. In 1962 Levins introduced a widely adopted graphical method for analyzing evolution towards an optimal combination of two quantitative traits, which are traded off. His approach explicitly excluded the possibility of density-dependent and frequency-dependent selection. Here we extend Levins method towards models, which include these selection regimes.

When selection is density-dependent and frequency-dependent, fitness landscapes change with population state. In the presence of frequency-dependence the invasion success of a given mutant not only depends on its traits but also on the densities of the various classes of residents. We derive the dynamics of evolution from the same types of curves Levins used: trade-off curves and fitness contours. However, fitness contours are not fixed but a function of the resident traits and we only consider those that divide the trait space into potentially successful mutants and mutants which are not able to invade. By means of the behavior of these two curves we can identify points where the fitness gradient is zero (`evolutionarily singular points') and assign them to the four possible types: Continuously Stable Strategies, Repellors, Branching Points and Gardens of Eden.

We show which a priori predictions about the evolutionary dynamics are possible when certain properties of the above mentioned curves are known. Furthermore we illustrate the framework by applying it to models from the literature.

In this talk we present a brief overview of our work in progress on the study of evolutionary processes emerging from an abstract logical form of self-reproduction. The model we use for our experiments pairs an evolving self-replicator (the "evoloop") with a dynamic environment (the "dissolver") in a deterministic cellular automata space. Populations in the resulting system are shown to exhibit speciation and long-term diversity leading to complex evolutionary behaviour. Using a new method for identification of loop species according to their configuration at birth, we present a detailed analysis of the genealogy in this CA space and show that it is strongly graph-based. Deviations from an exclusive tree-based hierarchy are moreover shown to be significant in understanding and describing the evolution of populations over time. We conclude with a visualization of graph-based representations for some illustrative test cases.

In Drosophila two distinct foraging strategies have been found to exist, namely 'rover' and 'sitter'. This behavioral polymorphism has been linked to the foraging gene ('for', or 'dg2'), which encodes for a cGMP-dependent protein kinase (PKG). The rover phenotype will move between separate foodpatches, while the sitter (having a lower PKG activity level) will not leave an existing foodpatch until it is exhausted.

A 2-layer synchronous cellular automata has been used to model the environment of the Drosophila larvae (yeast colonies growing on a fruit surface). One layer was used to model yeastgrowth, while the second contained a diffusing alcohol vapor generated by the fermenting activities of the yeast in the first layer.

Using the relative difference in alcohol concentration between their own location and the surrounding points, the larvae can find their way towards food particles (yeast). The two phenotypes were obtained by varying the sensitivity (or behavioral response) to a decline in alcohol concentration.

In this model of foraging behavior we show that the rover and sitter phenotype are alternate end attractors.

The advent of complete genome sequencing presents us with an unprecedented wealth of data. We will discuss how through comparative genome analysis, the availability of this data allows new insights into evolution. Apart from the intrinsic interest in these insights, we will show how this understanding allows genomes to be used for the prediction of gene function.

We study the specialisation of mixotrophs into separate autotrophs and heterotrophs. As a modelling approach we use the Dynamic Energy Budget theory (DEB), which is based on physiological rules for the organisms' metabolism. The mixotrophs can take up dissolved inorganic nutrients by autotrophic assimilation and detritus by heterotrophic assimilation. The organisms have a certain affinity for both assimilatory pathways. These affinities increase their (potential) assimilation rate, but they also entail certain costs. The two affinities (traits) together determine an organism's "strategy". This strategy is subject to evolution as at reproduction random mutations occur in either of the two affinities. An initially rare mutant with a slightly different strategy will be either better or less adjusted to the current state of the environment. As a result it will either die out, take over the whole population or live on in coexistence with the resident population of mixotrophs. Predictions of the evolutionary outcome of a system after a series of such mutations can be made with use of Adaptive Dynamics theory (AD). Two of the possible outcomes of evolution are the Continuously Stable Strategies (CSS's) and the branching points. The former are attracting strategies that can not be invaded by any mutant strategy. The population will evolve to such a strategy and once reached it will remain on it. The latter, the branching points, are attractors as well, but these can be invaded by mutants: At arriving the population will split up due to disruptive selection. The population will then change from consisting of a single strategy (monomorphic) to consisting of two strategies (polymorphic). Such a branching point provides an opportunity for our mixotrophic population to split up and specialise into separate autotrophs and heterotrophs. In order to find under what conditions this process of specialisation might occur, we study the evolutionary pathways and outcomes of the mixotroph model. Applying Adaptive Dynamics theory to this model is interesting also from a methodological point of view as the model's trait-space is two-dimensional (because of the two affinities that are involved) and because it is biologically more realistic than the toy-models that are used for developing the theory. Specialisation appeared to be not a common feature of the studied system, but is found only under specific conditions. These conditions only depend on the organism's mechanisms, not on system properties like total nutrient content. The DEB formulation facilitates the interpretation of the criteria for specialisation in terms of mechanisms underlying the organism's metabolism.

According to the received view, biologists distinguish themselves from other natural scientists by the kind of questions they ask about their subject matter. Whereas physicists and chemists limit themselves to how-questions, biologists are also interested in why-questions and answer these by means of evolutionary explanations. This view on the character of biology originates in Ernst Mayr's famous ``Cause and Effect in Biology'' (1961). It results in a dualistic image of biology as consisting of two ``largely separate fields:'' functional biology and evolutionary biology. Functional biology is seen as a reductionistic science that applies physics and chemistry in order to answer how-questions about organisms and their parts. Evolutionary biologiy is an autonomous brache of natural science that addresses why-questions by studying evolution. I argue that the source of biology's autonomy is the complexity of the life state, rather than the special character of the evolutionary process. Functional biologists do and should address why-questions in addition to how-questions. The functional explanations that answer this question are non-reductionist and a-historical in character.