This thesis presents two novel algorithms for the evolutionary optimisation of agent populations through divergent search of low dimensional phenotypic landscapes. As the eld of Evolutionary Robotics (ER) develops towards more complex domains, which often involve deception and uncertainty, the promotion of phenotypic diversity has become of increasing interest. Divergent exploration of the phenotypic feature space has been shown to avoid convergence towards local optima and to provide diverse sets of solutions to a given objective. Novelty Search (NS) and the more recent Multi-dimensional Archive of Phenotypic Elites (MAP-Elites), are two state of the art algorithms which utilise divergent phenotypic search. In this thesis, the individual merits and weaknesses of these algorithms are built upon in order to further develop the study of divergent phenotypic search within ER. An observation that the diverse range of individuals produced through the optimisation of novelty will likely contain solutions to multiple independent objectives is utilised to develop Multiple Assessment Directed Novelty Search (MADNS). The MADNS algorithm is introduced as an extension to NS for the simultaneous optimisation of multiple independent objectives, and is shown to become more e ective than NS as the size of the state space increases. The central contribution of this thesis is the introduction of a novel algorithm for rapid and thorough divergent search of low dimensional phenotypic landscapes. The Spatial, Hierarchical, Illuminated NeuroEvolution (SHINE) algorithm di ers from previous divergent search algorithms, in that it utilises a tree structure for the maintenance and selection of potential candidates. Unlike previous approaches, SHINE iteratively focusses upon sparsely visited areas of the phenotypic landscape without the computationally expensive distance comparison required by NS; rather, the sparseness of the area within the landscape where a potential solution resides is inferred through its depth within the tree. Experimental results in a range of domains show that SHINE signi cantly outperforms NS and MAP-Elites in both performance and exploration.

Background Recurrence and predictability of evolution are thought to reflect the correspondence between genomic and phenotypic dimensions of organisms, and the connectivity in deterministic networks within these dimensions. Direct examination of the correspondence between opportunities for diversification imbedded in such networks and realized diversity is illuminating, but is empirically challenging because both the deterministic networks and phenotypic diversity are modified in the course of evolution. Here we overcome this problem by directly comparing the structure of a “global” carotenoid network – comprising of all known enzymatic reactions among naturally occurring carotenoids – with the patterns of evolutionary diversification in carotenoid-producing metabolic networks utilized by birds. Results We found that phenotypic diversification in carotenoid networks across 250 species was closely associated with enzymatic connectivity of the underlying biochemical network – compounds with greater connectivity occurred the most frequently across species and were the hotspots of metabolic pathway diversification. In contrast, we found no evidence for diversification along the metabolic pathways, corroborating findings that the utilization of the global carotenoid network was not strongly influenced by history in avian evolution. Conclusions The finding that the diversification in species-specific carotenoid networks is qualitatively predictable from the connectivity of the underlying enzymatic network points to significant structural determinism in phenotypic evolution.

Experiments with microbes are a powerful tool for addressing general questions in evolutionary ecology. Microbial evolution is also interesting in its own right, and often clinically relevant. I have used experimental evolution of bacteria (Pseudomonas spp.) in controlled laboratory environments to investigate the role of environmental heterogeneity in the evolution of phenotypic diversity. Some of my results provide insight on general processes, while others are specific to bacteria. (1) I have shown that variation in resource supply affects the evolution of niche breadth in complex environments containing a range of available resources, leading to a peak in phenotypic diversity at intermediate levels. (2) I have found that resource availability also affects selection against redundant phenotypic characters, which is strongest when resources are scarce. (3) Using experiments with bacteria and their protozoan predators, I have found that selection for predator resistance varies with resource supply during a model adaptive radiation. (4) I have looked at the role of periodic bottlenecks in population size in the evolution of antibiotic-resistant bacteria. My results highlight the importance of biochemical constraints specific to different resistance mutations. (5) Finally, I have shown that bacterial adaptation to novel carbon substrates affects different growth parameters simultaneously, and that the same response is seen in environments that maintain different levels of phenotypic diversity. These findings emphasize the role of environmental heterogeneity in the evolution of phenotypic diversity, but also show how ecological and genetic factors can constrain adaptation to a given niche within a heterogeneous environment.

Most staple food crops were domesticated thousands of years ago through independent processes across different regions of the world. Studies of the history of such crops have been essential to our understanding of plant domestication as a process that started with the collection of wild material and continued with subsequent propagation, cultivation, and selection under human care. Domestication often involves a complex genetic structure with contributions from multiple founder populations, interspecific hybridization, chromosomal introgressions, and polyploidization events that occurred hundreds to thousands of years earlier. Such intricate origins complicate the systematic study of the sources of phenotypic variation. The analysis of recently domesticated, non-traditional, non-model species, such as Sinningia speciosa (Gesneriaceae), can expand the knowledge that we have on phenotypic variation under domestication, and help us to comprehend modern patterns of plant domestication and to broaden our understanding of the general trends. S. speciosa is commonly known as the 'florist's gloxinia', and it has been cultivated for 200 years as an ornamental houseplant. In our genomic study of S. speciosa, we examined an extensive diversity panel consisting of 115 individuals that included different species in the genus, wild representatives, and cultivated accessions, as well as 150 individuals from an F2 segregating population. Our analyses revealed that all of the domesticated varieties are derived from a single founder population that originated in or near the city of Rio de Janeiro in Brazil. We identified two loci associated with domesticated traits (flower symmetry and color) and did not detect any major hybridization or polyploidization events that could have contributed to the rapid increase in phenotypic diversity. Our findings, in conjunction with other features such as a small, low-complexity genome, ease of cultivation, and rapid generation time, makes this species an attractive model for the study of genomic variation under domestication. Basic research on non-model organisms with low economic importance is uncommon but necessary to understand the world from a broader perspective. In such cases, reduced representation approaches like Genotyping-by-Sequencing (GBS) are efficient low-cost alternatives to whole genome resequencing. However, most of these technologies are subject to patent protection, licensing processes, and fees that constrain genomic research for small non-profit research organizations. We have designed a protocol to construct reduced representation libraries from genomic DNA. Our approach, called Targeted Amplification of Scattered Sites (TASS), deviates from the traditional digestion-ligation-amplification process that is the subject of intellectual property that protects most current methods. Instead, TASS relies on 1) targeting and duplicating scattered regions in the genome by annealing and expanding long tail primers with short annealing sites, and 2) amplifying these regions using primers that are complementary to the added overhangs. At the moment GBS is more consistent and delivers more variants than TASS. However, we have established a foundation on which further optimization can produce an accessible, easy to implement, high-throughput genotyping approach. / Doctor of Philosophy / Most staple food crops were domesticated thousands of years ago through unrelated processes that were initiated across different regions of the world. Studies of the history of such crops have been essential to our understanding of plant domestication, a process that started with the collection of wild material and continued with subsequent propagation and cultivation under human care. Plant domestication has often involved a complex combination of ancestral lineages that encompass multiple populations, crosses with other species, and large DNA reorganizations that occurred hundreds to thousands of years earlier. Such intricate origins make the systematic study of plant domestication very challenging. The analysis of recently domesticated plants such as the 'florist's gloxinia' (Sinningia speciosa), can help us to better understand some of the changes that have occurred during domestication, as well as to comprehend modern patterns of plant domestication and to broaden our understanding of general trends. Florist's gloxinias are ornamental plants that have been cultivated during the last 200 years. In this study we examined 115 specimens, including wild and cultivated types of florist's gloxinias, as well as closely related species in Sinningia. We also constructed and evaluated an artificial population of 150 individuals from the cross of a wild and a cultivated form. Our analyses revealed that all of the domesticated varieties are descendants from a single wild population that originated in or near the city of Rio de Janeiro in Brazil. We also identified two regions of DNA that are responsible for the changes in flower shape and color, and crosses with other species did not introduce such alterations. Our findings, in conjunction with other features such as its small nuclear DNA content, the ease of cultivation indoors, and a rapid generation time, makes the florists' gloxinia an attractive crop to the study the effects of plant domestication. Research on organisms with low economic importance is uncommon but necessary to understand the world from a broader perspective. In such cases, analyzing the entire genetic information that is stored as DNA may be cost-prohibitive. Instead, approaches that sample small portions of DNA from each individual can be utilized. Most of these technologies are currently patented and subject to licensing processes and fees that limit their implementation by small non-profit research organizations. In this study we designed a protocol to sample small portions of DNA, similarly to existing techniques. However, our approach, called Targeted Amplification of Scattered Sites (TASS), employs a sampling process that deviates from the traditional patented procedure that is used in most current methods. At present, TASS is not as consistent and delivers less information than traditional approaches. However, we have established a foundation on which further optimization can produce an accessible and easy to implement technique.

Submitted by M?rden L?les (marden.inacio@ufvjm.edu.br) on 2016-07-22T18:27:43Z No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) D?bora_Ara?jo_de _Carvalho.pdf: 1646528 bytes, checksum: 9dc491b7402a28d39731f09e2ae99069 (MD5) / Approved for entry into archive by Rodrigo Martins Cruz (rodrigo.cruz@ufvjm.edu.br) on 2016-07-25T17:36:26Z (GMT) No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) D?bora_Ara?jo_de _Carvalho.pdf: 1646528 bytes, checksum: 9dc491b7402a28d39731f09e2ae99069 (MD5) / Made available in DSpace on 2016-07-25T17:36:26Z (GMT). No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) D?bora_Ara?jo_de _Carvalho.pdf: 1646528 bytes, checksum: 9dc491b7402a28d39731f09e2ae99069 (MD5) Previous issue date: 2016 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior (CAPES) / Canelas-preta s?o galinhas nativas brasileiras, encontradas no estado do Piau? e provavelmente em outros estados do Nordeste. Caracteriza-se por possuir carne de colora??o escura. Sua plumagem ? predominantemente preta, com varia??es de cores na regi?o do pesco?o (branco, dourado e preto) e a cor do dorso ? preta. S?o criadas em sistema tradicionais, a campo. S?o aves que apresentam pouca exig?ncia em manejo, aparentemente s?o r?sticas e resistentes ?s doen?as e parasitas. Objetivou-se caracterizar geneticamente com uso de doze loci de microssat?lites as galinhas nativas Canelas-Preta do estado do Piau? e, tamb?m, caracterizar e analisar a diversidade fenot?pica dessas aves com base em descritores fenot?picos. Para caracteriza??o gen?tica e fenot?pica foram utilizadas, respectivamente, 118 e 116 aves de tr?s munic?pios do estado. Para as an?lises gen?ticas foram utilizados 12 loci de microssat?lites e para as an?lises fenot?picas foram usados 32 descritores morfol?gicos, sendo 21 quantitativos e 11 qualitativos. Para o estudo gen?tico foram estimadas frequ?ncias al?licas em cada loco, heterozigosidades esperada (He) e observada (Ho), ocorr?ncia do equil?brio de Hardy-Weinberg, n?mero efetivo de popula??es e valores de PIC. Tamb?m foram realizadas as an?lises da estat?stica F pela an?lise do FIS (coeficiente de endogamia), AMOVA e componentes principais. Foi gerada uma matriz de dissimilaridade, gr?fico de dispers?o. A an?lise de estrutura populacional foi realizada usando o software STRUCTURE. Na caracteriza??o fenot?pica, os caracteres quantitativos foram submetidos a uma an?lise de vari?ncia. A an?lise estat?stica foi feita utilizando o m?todo dos quadrados m?nimos tendo sido realizadas an?lise da m?dia, m?nimo, m?ximo e coeficiente de varia??o de cada vari?vel e para cada n?cleo. A diversidade gen?tica foi obtida por meio da an?lise de agrupamento. As galinhas nativas Canelas-Preta, apresentaram elevada variabilidade gen?tica para os loci analisados, o que mostra a conserva??o desse material gen?tico. Foi poss?vel verificar a exist?ncia de alta variabilidade gen?tica intrapopulacional, o que indica que os n?cleos foram formados com suficiente variabilidade gen?tica (efeito fundador). Essa elevada diferencia??o gen?tica dentro das popula??es somada a baixa diferencia??o entre as popula??es, permite afirmar que os indiv?duos analisados pertencem a ?nico grupo gen?tico. As galinhas Canelas-Preta apresentam caracter?sticas de aves nativas, uma vez que mostraram semelhan?a nas distribui??es dos descritores qualitativos entre os n?cleos amostrados e varia??o na frequ?ncia desses descritores dentro de cada n?cleo, isso a caracteriza como ra?a e fenotipicamente estruturada, com padr?es uniformes. Possuem caracter?sticas fenot?picas quantitativas de elevado, m?dio e pequeno coeficiente de varia??o, que revela a riqueza genica da ra?a e as qualifica para programas de conserva??o, utiliza??o e melhoramento dos recursos gen?ticos. / Disserta??o (Mestrado) ? Programa de P?s-Gradua??o em Zootecnia, Universidade Federal dos Vales do Jequitinhonha e Mucuri, 2016. / ABSTRACT Canelas-preta are Brazilian native chickens, found in the state of Piau? and probably in other states of the Northeast, characterized by having dark colored meat. The plumage is predominantly black with color variations in the hackle (white, gold and black) and the back color is black. These birds are raised in extensive system, allowing little demand in management, being apparently rustics and resistant to diseases and parasites. This study aimed to characterize genetically Canelas-Preta native chickens of the Piau? state by using twelve microsatellite loci and also to characterize and analyze the phenotypic diversity of these birds based on phenotypic descriptors. For genetic and phenotypic characterization were used, respectively, 118 and 116 birds from three towns of the state. For the genetic analyzes were used 12 microsatellite loci and for phenotypic analyzes were used 32 morphological descriptors: 21 quantitative and 11 qualitative. For the genetic study, allele frequencies at each locus, expected heterozygosity (He) and observed (Ho), the occurrence of Hardy-Weinberg equilibrium, effective number of populations and PIC values were estimated. Furthermore, also it have been carried out F statistical analyzes by the analysis of FIS (inbreeding coefficient), AMOVA and main components. One dissimilarity matrix and a scatter plot were generated. The population structure analysis was performed using the STRUCTURE software. In the phenotypic characterization, quantitative characters were subjected to an analysis of variance. The statistical analysis was done by using the method of least squares, the analysis of average, minimum, maximum and coefficient of variation of each variable and for each core was performed. Genetic diversity was obtained by cluster analysis. The Canelas-Preta native chickens presented a high genetic variability for the loci analyzed, which shows the conservation of this genetic material. It was possible to verify the existence of high intrapopulation genetic variability, indicating that the nuclei were formed with sufficient genetic variability (founder effect). This high genetic differentiation within populations coupled with lower differentiation among populations, allows affirming that individuals analyzed belong to single genetic group. The Canelas-Preta native chicken have characteristics of native birds, once there was demonstration of similarity in the distributions of qualitative descriptors between sampled cores and variation in the frequency of those descriptors within each core, so this characterize the Canelas-Preta as a phenotypically structured breed with uniform patterns. This breed has quantitative phenotypic characteristics of high, medium and low coefficient of variation, which reveals its genetic wealth and qualify them for conservation programs, use and improvement of genetic resources.

Knowledge of the genetic variation of crop collections is essential for their efficient use in plant breeding programs. The Ethiopian Highland Maize Germplasm Collection Mission was launched throughout the highlands of Ethiopia in 1998 and 287 traditional maize accessions were collected from farmers’ fields. To date, no information was available on the morphological and genetic diversity in this important collection. Various molecular marker techniques and quantitative genetics approaches were applied to accurately unravel the extent of phenotypic and genetic diversity, to study patterns of morphological and molecular variation and to determine association of molecular markers with quantitative trait variation, with the view of designing a sound breeding program and management strategy for maize in the highlands of Ethiopia. The morphological study confirmed that traditional Ethiopian highland maize accessions contain large amounts of variation for agro-morphological traits. The broad trait diversity observed among the accessions suggested ample opportunities for the genetic improvement of the crop through selection directly from the accessions and/ or the development of inbred lines for a future hybrid program. Selection practices followed by local farmers are mostly consistent within agroecology and gave rise to morphologically distinct maize accessions in different agroecologies. This underscores the importance of considering farmers’ knowledge of diversity in the collection and evaluation of local accessions. The results of amplified fragment length polymorphism (AFLP) and microsatellite or simple sequence repeat (SSR) marker analyses showed that bulking leaf samples from 15 individual plants per out-bred accession is an effective means of producing representative profiles of individual plants, thereby reducing the cost of DNA extraction and subsequent marker analysis of open-pollinated varieties. Cluster analyses based on AFLP and SSR data showed that most of the accessions collected from the Northern agroecology were genetically distinct from the Western and Southern accessions suggesting that differentiation for adaptive traits for drought conditions may have occurred in the Northern accessions. However, there was very little genetic differentiation between the Western and Southern accessions suggesting gene flow between the two agroecologies and recent introduction of similar improved varieties in these agroecoogies . In both marker systems, high mean genetic diversity was observed among the traditional Ethiopian highland maize accessions. This is possibly due to (i) the continuous introduction of maize from abroad by different organizations; (ii) genetic variation generated through farmers management practices; and (iii) the presence of different environmental conditions in the highlands of Ethiopia to which local landraces may have been adapted. The correlation between the morphological dissimilarity matrix and the matrices of genetic dissimilarity based on SSR and AFLP markers were 0.43 and 0.39, respectively (p = 0.001 in both cases). The correlation between SSR and AFLP dissimilarity matrices was 0.67 (p = 0.001). These significant correlations indicate that the three independent sets of data likely reflect the same pattern of genetic diversity, and validate the use of the data to calculate the different diversity statistics for Ethiopian highland maize accessions. From this study, three groups of maize accessions with distinctive genetic profiles and morphological traits were identified that will be useful for future collection, conservation and breeding programs of maize for the highlands of Ethiopia. A pilot association study using SSR markers and quantitative trait variation indicated that molecular markers could be useful to identify genetic factors controlling earliness, tallness, grain yield and associated traits, which could be exploited by various breeding schemes. The analytical tools outlined in this dissertation can be a useful tool in managing genetic variation of open-pollinated crops and will aid in the conservation of unique genetic diversity. Production stability and global food security are linked to the conservation and exploitation of worldwide genetic resources and this research attempts to add to that body of knowledge. Copyright 2005, University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria. Please cite as follows: Beyene, YA 2005, Genetic analysis of traditional Ethiopian Highland Maize (Zea Mays l.) using molecular markers and morphological traits : implication for breeding and conservation, PhD thesis, University of Pretoria, Pretoria, viewed yymmdd < http://upetd.up.ac.za/thesis/available/etd-02212006-112610 / > / Thesis (PhD (Genetics))--University of Pretoria, 2005. / Genetics / unrestricted

The canine transmissible venereal tumour (CTVT) is a contagious cancer that is naturally transmitted between dogs by the allogeneic transfer of living cancer cells during coitus. CTVT first arose several thousand years ago and has been reported in dog populations worldwide. The goals of this Thesis were (1) to gain further understanding of CTVT distribution patterns and prevalence around the world, (2) to use genetics to trace the historical spread of CTVT and (3) to map the genetic as well as phenotypic diversity of CTVT tumours around the world. To understand the distribution patterns of CTVT, I obtained information from 645 veterinarians and animal health workers in 109 countries, and generated a snapshot of the locations in which this disease is found. Additionally, as preparation for further genetic analysis, I collected samples from over one thousand CTVT cases from more than 50 countries, optimised methods for high-throughput DNA extraction and quantification and optimised a qPCR-based assay for CTVT diagnosis and host contamination detection. With the goal of tracing the historical spread of CTVT and learning about the genetic diversity of this disease, I sequenced complete mitochondrial genomes of 449 CTVT tumours and their matched hosts. The analysis of the CTVT mitochondrial diversity revealed that CTVT has captured mitochondrial DNA (mtDNA) through horizontal transfer events at least five times during the history of the lineage, delineating five tumour clades. CTVT appears to have spread rapidly around the world within the last 2,000 years, perhaps transported by dogs travelling along historic maritime trade routes. This work indicated that negative selection has operated to prevent accumulation of deleterious mutations in captured mtDNA, and that recombination has caused occasional mtDNA re-assortment. A histology-based screen of CTVT clades did not show any significant phenotypic differences between groups. In order to determine how the five mtDNA clades relate to each other, I analysed data from 539 CTVT exomes. This revealed that a single canine mtDNA haplogroup has recurrently and recently undergone multiple horizontal transfer events. Analysis of this haplotype highlighted a number of candidate genetic variants which may be conferring a selective advantage to this haplotype in CTVT, possibly by influencing mtDNA transcription or replication. Overall, genetic and phenotypic analysis of CTVT tumours from across the globe has broadened our understanding of CTVT diversity, and provided important insights into the biology of a unique transmissible cancer.