Thesis (Ph.D.(Sustainable Ecological Management))--University of Pretoria, 2000. / Includes bibliographical references.

Most eukaryotes harbor a diverse community of parasitic, mutualistic, and commensal microbial symbionts. Although the diversity of these microbial symbiotic communities has recently drawn considerable attention, theory regarding the evolution of interactions among symbionts and with the host is still in its nascent stages. Here we evaluate the role of interactions among coinfecting symbionts in the evolution of symbiont virulence toward the host. To do so, we place the virulence-transmission trade-off into a community context and model the evolution of symbiont trophic modes along the continuum from parasitism (virulence) to mutualism (negative virulence). We establish a framework for studying multiple infections of a host by the same symbiont species and coinfection by multiple species, using a concept of shared costs, wherein the negative consequences of virulence (or harm) toward the host are shared among symbionts. Our results show that mutualism can be maintained under infection by multiple symbionts when shared costs are sufficiently low, while greater virulence and parasitism toward the host are more likely when shared costs are high. Last, for coinfection by more than one species, we show that if the presence of a mutualist ameliorates some of the costs of pathogen virulence, then the symbiotic community may more often evolve to a more commensal state and maintain mutualisms.

Virtually all interacting species (such as hosts and parasites) are embedded within diverse communities. However, evolutionary interactions are typically considered in a pairwise species framework. Although coevolutionary theory suggests that multiple species interactions may provide greater opportunities for diversification, the impacts of community diversity on coevolution have not been directly tested. In this thesis I synthesize the findings from recent experimental work to assess the effects of increased species diversity on the patterns and processes of host and parasite evolution. I then investigate the effects of parasite diversity on host-parasite population dynamics and evolution using the pathogen Pseudomonas aeruginosa and five lytic bacteriophage parasites in a brief evolution experiment. Parasite diversity was manipulated by assembling phage communities with different number of species. Phage communities suppressed host populations more rapidly but also showed reduced phage density, likely due to inter-phage competition. The evolution of resistance allowed rapid bacterial recovery that was greater in magnitude with increases in phage diversity. These results were then followed up via longer term experimental coevolution of the same host and parasite communities. Here the data showed that greater parasite diversity accelerates coevolutionary arms races and drives more diversification among lineages. Coevolution between hosts and parasite communities drove more successive increases in host resistance coupled with increasingly frequent selective sweeps at the genomic level. Consistent with this, the most rapidly evolving host genes under coevolution with parasite communities were those involved in various host resistance strategies. These results demonstrate, at phenotypic and genomic levels, how areas of high community diversity may be hotspots for rapid evolution in interacting, antagonistic species. Finally, In the face of escalating antibiotic resistance, there is now an urgent need to develop alternative antimicrobials, these results may be relevant to the application of phages as therapeutics and they are discussed in that context.

Under the Dobzhansky-Muller model, postzygotic isolation results from incompatibility between interacting genes. Evidence points to regulatory networks as a rich source of incompatibilities that impact hybrid fitness. Pleiotropy is a natural feature of regulatory networks because regulatory elements generally have multiple targets. Both pleiotropy and hybrid incompatibility arise due to genetic interactions; therefore we can expect an intimate association between them. In the following chapters, I investigate the relationship between pleiotropy and hybrid incompatibility in the context of regulatory networks. In chapter one, I extend a general network-based study of hybrid incompatibility by incorporating a sequence-based thermodynamic model of transcriptional regulation. In the absence of pleiotropy, hybrid misregulation of a positively selected trait evolves quickly as a consequence of non-recognition or spurious binding in regulatory interactions across species boundaries. In a conserved trait, hybrid incompatibility evolves much slower as a product of compensatory drift. In chapter two, I show that pleiotropy can promote or constrain the evolution of hybrid incompatibility in a regulatory network depending on its fitness landscape, which emerges from the thermodynamic properties of molecular binding. Pleiotropy may promote hybrid incompatibility in accordance with the "selection, pleiotropy, and compensation model" of evolution, in which compensation for the pleiotropic side-effects of adaptation accelerates incompatibility in conserved traits. Pleiotropy can limit the evolution of hybrid incompatibility by constraining change in trans-acting regulatory elements in favor of adaptation at less pleiotropic downstream cis-regulatory targets. Without change in both interactors, incompatibility does not occur under the Dobzhansky-Muller model. In chapter three, I evaluate the hypothesis that pleiotropy facilitates the onset of hybrid incompatibility under antagonistic coevolution, an ubiquitous and persistent source of natural selection. When infectivity and resistance in a host-parasite system are determined epistatically by network interactions, reciprocal selective pressure results in a genotypic chase. This causes pleiotropic mutations to accumulate and be compensated over time, producing intrinsic hybrid incompatibility in both species independent of local adaptation. Thus, cyclical antagonistic coevolution eventually overcomes constraint on pleiotropic loci, facilitating the evolution of regulatory incompatibilities commonly observed in hybrids.

As pressões seletivas impostas por interações ecológicas são uma das forças que moldam a adaptação por seleção natural em populações. Entre os resultados possíveis das pressões seletivas impostas por interações está a coevolução, isto é, mudanças evolutivas recíprocas que ocorrem nas populações das espécies que interagem. Um dos principais desafios para a ecologia evolutiva é entender se e como o processo coevolutivo ocorre quando espécies interagem com muitas outras espécies formando redes de interações. Nesta tese desenvolvi, com a ajuda de colaboradores, modelos que descrevem a coevolução entre espécies que interagem de forma antagonista. Interações antagonistas são interações ecológicas interespecíficas que resultam em consequências negativas para a aptidão de indivíduos de uma das espécies envolvidas e positivas para indivíduos da outra espécie. Busquei uma melhor compreensão sobre os mecanismos ecológicos e evolutivos responsáveis pela formação, manutenção e evolução das redes de interação antagonista. Em primeiro lugar, encontrei que a assimetria da seleção influenciou a dinâmica evolutiva em antagonismos. A dinâmica coevolutiva gerou corridas armamentistas quando a intensidade da seleção foi maior sobre as vítimas do que sobre os exploradores. Por outro lado, os valores dos fenótipos flutuaram quando a intensidade da seleção foi maior sobre os exploradores do que sobre as vítimas. No entanto, a dinâmica coevolutiva dependeu da estrutura das redes formadas por antagonistas. Redes aninhadas favoreceram a evolução de resistência em vítimas atacadas por exploradores generalistas. A dinâmica evolutiva também reorganizou as redes de interação e, especialmente em cenários nos quais a seleção favoreceu forte acoplamento fenotípico, formou módulos de espécies interagentes. Em segundo lugar, encontrei que regras de interação baseadas no acoplamento fenotípico ou em barreiras fenotípicas reproduziram a estrutura de redes antagonistas empíricas, mas as duas relações funcionais entre fenótipos e aptidão tenderam a subestimar o aninhamento e superestimar a modularidade das redes empíricas. No entanto, a evolução das características foi diferentemente moldada por essas relações funcionais, sendo mais flutuante no modelo de acoplamento fenotípico e mais direcional no modelo de barreiras fenotípicas. Portanto, a coevolução mediada por diferentes relações funcionais resultou em diferentes dinâmicas coevolutivas mas não teve impacto sobre a organização das redes de interação antagonistas. Em terceiro lugar, estudei como variações nas abundâncias e nos fenótipos estão relacionadas e encontrei que a coevolução rápida mediada por forte pressões seletivas impostas por interações ecológicas pode resultar em uma baixa variabilidade nas abundâncias das populações e alta variabilidade fenotípica. Em contraste, em cenários nos quais a seleção imposta por interações é fraca, encontrei uma alta variabilidade nos tamanhos populacionais e baixa variabilidade fenotípica. Portanto, a rápida resposta evolutiva reduziu as flutuações nos tamanhos populacionais, reduzindo extinções devido a flutuações demográficas. Porém, este resultado foi influenciado pela estrutura da rede: a modularidade aumentou a estabilidade das interações enquanto que o aninhamento esteve associado a maior flutuação demográfica. Por fim, estudei espalhamento de um parasita que infecta diferentes espécies de hospedeiros e que pode ser transmitido por meio da predação de um hospedeiro infectado ou por meio de vetores biológicos. Combinei as diferentes redes antagonistas formadas a partir das interações mediadas por cada mecanismo de transmissão em uma rede de interação múltipla espacialmente explícita. Por meio de um modelo matemático, obtive que a transmissão do parasita é maximizada quando ambos os mecanismos de transmissão são considerados ao mesmo tempo e quando os processos ocorrem com probabilidade semelhante. A análise da cartografia da rede múltipla aliada a simulações de imunização de diferentes tipos de hospedeiros mostraram que a estrutura da rede múltipla pode indicar o papel que cada espécie de hospedeiro desempenha na transmissão do parasita em um determinado ecossistema / Mutualisms are interactions in which organisms of different species exploit each other with net benefits for both interacting individuals. Multispecific mutualistic system can be depicted as interaction networks, such as those formed by plant-pollinator interactions, dispersal systems, species interacting in cleaning stations in reef environments, protective ants in plants, müllerian mimicry, and nitrogen fixing bacteria on the roots of plants. Mutualistic interaction is subject to cheating by individuals who, by means of a diversity of behavioral strategies, achieve the benefit provided by the partner offering nothing or few in return. However, the mutualistic interactions persist despite the existence of cheaters. In this work I show that the parasites of mutualistic interactions increase the resilience of mutualistic networks to disturbances in nested networks, typically found in species-rich mutualisms. Therefore the joint effect of cheating, structure and dynamics of mutualistic networks have implications for how biodiversity is maintained. I subsequently study the conditions under which tubular flowers, which suffer stronger damages when interacting with nectar robbers, can coexist with planar flowers, pollinators, and robbers through indirect effects of cheating on their reproductive success. The theft of nectar may increase the success of a plant if its interactions with robbers generate higher degrees of cross-pollination, thus increasing the reproductive success of plants that interact with both floral visitors. This study suggests a new source of continued cooperation and diversity strategies through non-linear effects of the interactions between different strategies. Finally, I study how local interactions can promote the prevalence of mimic (the cheaters) in a given population in the absence of their models. I found that prey interacting locally may favor the predominance of mimic preys and avoid predators that, after a few generations and under a non-random distribution of individuals in space, can further strengthen this unexpected effect allopatry of the mimic and its model

Understanding how changes in functional requirements of the cell select for changes in protein sequence and structure is a fundamental challenge in molecular evolution. This dissertation delineates some of the underlying evolutionary forces using as a model system, the Haloalkanoate Dehalogenase Superfamily (HADSF). HADSF members have unique cap-core architecture with the Rossmann-fold core domain accessorized by variable cap domain insertions (delineated by length, topology, and point of insertion). To identify the boundaries of variable domain insertions in protein sequences, I have developed a comprehensive computational strategy (CapPredictor or CP) using a novel sequence alignment algorithm in conjunction with a structure-guided sequence profile. Analysis of more than 40,000 HADSF sequences led to the following observations: (i) cap-type classes exhibit similar distributions across different phyla, indicating existence of all cap-types in the last universal common ancestor, and (ii) comparative analysis of the predicted cap-type and functional diversity indicated that cap-type does not dictate the divergence of substrate recognition and chemical pathway, and hence biological function. By analyzing a unique dataset of core- and cap-domain-only protein structures, I investigated the consequences of the accessory cap domain on the sequence-structure relationship of the core domain. The relationship between sequence and structure divergence in the core fold was shown to be monotonic and independent of the corresponding cap type. However, core domains with the same cap type bore a greater similarity than the core domains with different cap types, suggesting coevolution of the cap and core domains. Remarkably, a few degrees of freedom are needed to describe the structural diversity in the Rossmann fold accounting for the majority of the observed structural variance. Finally, I examined the location and role of conserved residue positions and co-evolving residue pairs in the core domain in the context of the cap domain. Positions critical for function were conserved while non-conserved positions mapped to highly mobile regions. Notably, we found exponential dependence of co-variance on inter-residue distance. Collectively, these novel algorithms and analyses contribute to an improved understanding of enzyme evolution, especially in the context of the use of domain insertions to expand substrate specificity and chemical mechanism.

The vaginal microbiome plays an important role in reproductive health and pregnancy. It has coevolved with humans and has direct effects on reproductive success, rendering selective pressure more pronounced at this site. Herein, we probe coevolution of the vaginal microbiome using a systems-level approach. In Chapter 2, we examine the evolutionary trajectory of two vaginal Veillonellaceae phylotypes evolved from an ancestral gastrointestinal lineage to inhabit the vaginal niche. We present evidence of their divergence and subniche specification and describe their differential associations with vaginal infection and pregnancy. In Chapter 3, we identify ten bacterial taxa, predicted to contribute to the underlying pathology of the sexually transmitted infection trichomoniasis. This ‘pathogroup’, which has undergone conditional differentiation to thrive in the presence of T. vaginalis, includes previously undescribed organisms and putative symbionts. Lastly in Chapter 4, we present the first characterization of BspA proteins, multi-modal virulence factors, in the vaginal microbiome and provide evidence of their extensive horizontal transfer across diverse microbial lineages. We use homology modeling to demonstrate conservation of structural and functional characteristics of these proteins between diverse bacterial taxa and identify structural variants, potentially indicative of subtypes. These findings further our understanding of the contributions of individual bacterial species, bacterial communities and virulence determinants in the health and disease. Furthermore, they lay the groundwork for future work characterizing coevolution of the human vaginal microbiome. These systems-level approaches will facilitate synergy between broad and reductive approaches and inform strategies for modulation of the microbiome and development of more effective therapeutics.

<p>Artificial intelligence plays an increasingly important role in modern computer games. As the complexity of the games increase, so does the complexity of the AI.</p><p>The aim of this dissertation is to investigate how AI for a turnbased computer game can coevolve into playing smarter by combining genetic algorithms with neural networks and using a reinforcement learning regime.</p><p>The results have shown that a coevolved AI can reach a high performance in this kind of turnbased strategy games. It also shows that how the data is coded and decoded and which strategy that is used plays a very big role in the final results</p>

This thesis investigates the use of symbiosis as an evolutionary metaphor for problem decomposition using Genetic Programming. It begins by drawing a connection between lateral problem decomposition, in which peers with similar capabilities coordinate their actions, and vertical problem decomposition, whereby solution subcomponents are organized into increasingly complex units of organization. Furthermore, the two types of problem decomposition are associated respectively with context learning and layered learning. The thesis then proposes the Symbiotic Bid-Based framework modeled after a three-staged process of symbiosis abstracted from biological evolution. As such, it is argued, the approach has the capacity for both types of problem decomposition. Three principles capture the essence of the proposed framework. First, a bid-based approach to context learning is used to separate the issues of `what to do' and `when to do it'. Whereas the former issue refers to the problem-specific actions, e.g., class label predictions, the latter refers to a bidding behaviour that identifies a set of problem conditions. In this work, Genetic Programming is used to evolve the bids casting the method in a non-traditional role as programs no longer represent complete solutions. Second, the proposed framework relies on symbiosis as the primary mechanism of inheritance driving evolution, where this is in contrast to the crossover operator often encountered in Evolutionary Computation. Under this evolutionary metaphor, a set of symbionts, each representing a solution subcomponent in terms of a bid-action pair, is compartmentalized inside a host. Communication between symbionts is realized through their collective bidding behaviour, thus, their cooperation is directly supported by the bid-based approach to context learning. Third, assuming that challenging tasks where problem decomposition is likely to play a key role will often involve large state spaces, the proposed framework includes a dynamic evaluation function that explicitly models the interaction between candidate solutions and training cases. As such, the computational overhead incurred during training under the proposed framework does not depend on the size of the problem state space. An approach to model building, the Symbiotic Bid-Based framework is first evaluated on a set of real-world classification problems which include problems with multi-class labels, unbalanced distributions, and large attribute counts. The evaluation includes a comparison against Support Vector Machines and AdaBoost. Under temporal sequence learning, the proposed framework is evaluated on the truck reversal and Rubik's Cube tasks, and in the former case, it is compared with the Neuroevolution of Augmenting Topologies algorithm. Under both problems, it is demonstrated that the increased capacity for problem decomposition under the proposed approach results in improved performance, with solutions employing vertical problem decomposition under temporal sequence learning proving to be especially effective.

Significant phylogenetic codivergence between plant or animal hosts (H) and their symbionts or parasites (P) indicate the importance of their interactions on evolutionary time scales. However, valid and realistic methods to test for codivergence are not fully developed. One of the systems where possible codivergence has been of interest involves the large subfamily of temperate grasses (Pooideae) and their endophytic fungi (epichloae). Here we introduce the MRCALink (most-recent-common-ancestor link) method and use it to investigate the possibility of grass-epichloё codivergence. MRCALink applied to ultrametric H and P trees identifies all corresponding nodes for pairwise comparisons of MRCA ages. The result is compared to the space of random H and Ptree pairs estimated by a Monte Carlo method. Compared to tree reconciliation the method is less dependent on tree topologies (which often can be misleading), and it crucially improves on phylogeny-independent methods such as ParaFit or the Mantel test by eliminating an extreme (but previously unrecognized) distortion of node-pair sampling. Analysis of 26 grass species-epichloё species symbioses did not reject random association of H and P MRCA ages. However, when five obvious host jumps were removed the analysis significantly rejected random association and supported grass-endophyte codivergence. Interestingly, early cladogenesis events in the Pooideae corresponded to early cladogenesis events in epichloae, suggesting concomitant origins of this grass subfamily and its remarkable group of symbionts.