La recherche développée dans le cadre de cette thèse aborde différents aspects relevant de l’analyse statistique de données. Dans un premier temps, une analyse de trois indices d’associations entre deux tableaux de données est développée. Par la suite, des stratégies d’analyse liées à la standardisation de tableaux de données avec des applications en analyse en composantes principales (ACP) et en régression, notamment la régression PLS sont présentées. La première stratégie consiste à proposer une standardisation continuum des variables. Une standardisation plus générale est aussi abordée consistant à réduire de manière graduelle non seulement les variances des variables mais également les corrélations entre ces variables. De là, une approche continuum de régression a été élaborée regroupant l’analyse des redondances et la régression PLS. Par ailleurs, cette dernière standardisation a inspiré une démarche de régression biaisée dans le cadre de régression linéaire multiple. Les propriétés d’une telle démarche sont étudiées et les résultats sont comparés à ceux de la régression Ridge. Dans le cadre de l’analyse de plusieurs tableaux de données, une extension de la méthode ComDim pour la situation de K+1 tableaux est développée. Les propriétés de cette méthode, appelée P-ComDim, sont étudiées et comparées à celles de Multiblock PLS. Enfin, la situation où il s’agit d’évaluer l’effet de plusieurs facteurs sur des données multivariées est considérée et une nouvelle stratégie d’analyse est proposée. / The research developed in this thesis deals with several statistical aspects for analyzing datasets. Firstly, investigations of the properties of several association indices commonly used by practitioners are undergone. Secondly, different strategies related to the standardization of the datasets with application to principal component analysis (PCA) and regression, especially PLS-regression were developed. The first strategy consists of a continuum standardization of the variables. The interest of such standardization in PCA and PLS-regression is emphasized.A more general standardization is also discussed which consists in reducing gradually not only the variances of the variables but also their correlations. Thereafter, a continuum approach was developed combining Redundancy Analysis and PLS-regression. Moreover, this new standardization inspired a biased regression model in multiple linear regression. Properties related to this approach are studied and the results are compared on the basis of case studies with those of Ridge regression. In the context of the analysis of several datasets in an exploratory perspective, the method called ComDim, has certainly raised interest among practitioners. An extension of this method for the analysis of K+1 datasets was developed. Properties related to this method, called P-ComDim, are studied and compared to Multiblock PLS. Finally, for the analysis of datasets depending on several factors, a new approach based on PLS regression is proposed.

Cette thèse propose un test empirique du modèle d’attribution du leadership basé sur l’observation de dyades leader-subordonné au Sénégal et en France. Ce modèle se concentre sur des construits essentiels comme la complexité attributionnelle et l’exactitude des attributions des leaders, les comportements qui en résultent, les variables médiatrices comme la satisfaction des subordonnés, les stratégies correctives de la performance et les variables de résultat comme les perceptions du leadership. Le modèle est testé à l’aide de régressions hiérarchiques des effets directs et des effets médiateurs des attributions biaisées. Notre intention est de tester les relations à partir d’un échantillon d’auditeurs en formation continue de l’enseignement supérieur (au Sénégal et en France) et de déterminer selon notre modèle et nos hypothèses de recherche, quels construits contribuent le plus à expliquer la complexité attributionnelle des leaders. Cequi nous amène à étudier l’influence directe de l’exactitude des attributions, l’effet des stratégies correctives et de la satisfaction des subordonnés sur la perception du leadership. Nous testerons aussi le rôle des attributions biaisées sur les deux variables médiatrices. Les hypothèses sont fortement corroborées. La complexité attributionnelle des leaders est liée à l’exactitude de leurs attributions, telle qu’elle est perçue par leurs subordonnés. Les stratégies de correction de la performance élaborées par les leaders se sont révélées liées à des attributions exactes, puis à des variables clés de résultats. Nous analysons les résultats ainsi que les limites et les orientations futures de la recherche. Les contributionsmanagériales sont doubles : d'une part, de prendre en compte dans le processus de recrutement des leaders leur aptitude à adopter une approche attributionnelle complexe et, d'autre part, les former afin de mieux accompagner leurs subordonnés. / This thesis proposes an empirical test of the leadership attribution model based on the observation of leader-subordinate dyads in Senegal and France. This model focuses on key constructs such as attributional complexity and leader attribution accuracy, resulting behaviors, mediating variables such as subordinate satisfaction, performance corrective strategies, and outcome variables such as leadership perceptions. The model is tested using hierarchical regressions of direct effects and mediating effects of biased allocations. Our intention is to test the relationships from a sample of auditors in continuingeducation of higher education (in Senegal and France) and to determine according to our model and our research hypotheses, which constructs contribute the most to explain the attributional complexity of the leaders. This leads us to study the direct influence of the accuracy of the attributions, the effect of the corrective strategies and the satisfaction of the subordinates on the perception of the leadership. We will also test the role of biased assignments on the two mediating variables. Hypotheses are strongly corroborated. The attributional complexity of leaders is linked to the accuracy of their attributions, as perceived by their subordinates. The performance correction strategies developed by leaders have beenlinked to accurate attribution and then to key outcome variables. We analyze the results as well as the limits and future directions of the research. The managerial contributions are twofold: on the one hand, to take into account in the process of recruiting leaders their ability to adopt a complex attributional approach and, on the other hand, to train them to better support their subordinates.

從一九九0年代開始，由於微電子、電腦、衛星通信、網際網路、光學纖維等資訊科技的發展，不僅造成社會生活的變遷，也帶動了國家競爭的新形勢。 在這樣的時代環境下，資訊知識發展與科技運用成為經濟發展成功的關鍵，唯有促使科技出現重大突破，才能提昇產業的競爭力。 在學術界，討論資訊科技的文獻上出現了生產力矛盾(productivity paradox)的說法，即他們發現使用資訊科技對於生產力的提升出現了不顯著甚至是負的結果。針對此，後續有許多學者提出各種解釋生產力矛盾的原因，諸如過度投資(overinvestment)、衡量錯誤 (mismeasurement)等問題。 因此，本研究以生產力指標中的總要素生產力為衡量對象，檢視資訊科技對台灣製造業是否具有矛盾的情形，並觀察當技術革新時，對人力資本的需求變化，資料期間為1995-2002年，採用的計量模型為揉合資料（pooling data）模式。 根據實證結果，自動化設備的應用出現生產力矛盾(productivity paradox)的情形，而人力資本對總要素生產力的貢獻則為顯著的正值，若觀察投入自動化生產設備對人力資本的需求，則發現有資本與技術互補（capital-skilled complementarity）的情形。因此推論：單獨增加自動化設備並無法提升總要素生產力，唯有透過自動化設備投資與人力資本的配合，兩者互補之下，以提升台灣製造業生產力。

Black grouse (Tetrao tetrix) is a bird species with a lek mating system found in the Palearctic boreal taiga. It is assumed that it has a continuous distribution along Scandinavia and Siberia, whereas in Central Europe it has declined during the last decades. The primary objective of this thesis was to obtain a deeper understanding of the history, systematic classification and the genetic structure of black grouse on different geographical scales using microsatellites and control region mtDNA sequences (CR). I determined how much the mating system, habitat fragmentation and historical population processes have influenced the partitioning of genetic diversity in this species. Phylogeographical results are consistent with a demographic population expansion, and the patterns of postglacial dispersal suggest that a glacial refugium was located somewhere in central Asia, and from there black grouse spread out to Europe following the retreat of glacial ice sheets. I suggest that the two European black grouse subspecies, T. t. Tetrix and T. t. britannicus correspond to only one subspecies: T. t. tetrix, and that this lineage has diverged from T.t. viridanus, a subspecies found in Kazakhstan. The British population is significantly divergent from the remaining Eurasian samples for microsatellites but it is not for mtDNA. Therefore, they should regard as a separate Management Unit and not as a subspecies. Furthermore, British black grouse occur in three independent genetic units, corresponding to Wales, northern England/southern Scotland and northern Scotland. There was also genetic structure within Sweden. Habitat fragmentation is the main cause of population genetic structure in southern Swedish black grouse. In contrast, low levels of genetic differentiation and high connectivity were found in northern Sweden due to female-biased dispersal. On a finer geographical scale, I found genetic differences between leks due to a mixture of related and unrelated individuals within leks. However, mean relatedness values hardly differed from zero. Some leks were similar to one another and I interpret this as a result of variation in local reproductive success and philopatry. These factors would cause genetic structuring but this by itself would not reveal that kin selection is operating within black grouse leks.

This study provides a historical-educational evaluation of the teaching of vocational subjects at the Indian secondary school. The Indian secondary school curriculum is examined and evaluated and recommendations are made based on these findings for the vocational Africa. education future development of in contemporary South The study is structured in the following manner: * Chapter One Examines the concept of vocational education. * Chapter Two Traces the development of Indian education since 1860. * Chapter Three Retrospective look at the development of vocational education in the Indian education system. * Chapter Four Provides an exposition of the vocational education curriculum at the Indian secondary school. * Chapter Five Evaluates the vocational education curriculum. * Chapter Six Recommendations for the future of vocational education in South Africa. / Educational Studies / M. Ed. (History of Education)

Orientador: John Fredy Franco Baquero / Resumo: Nesta dissertação é proposta a aplicação do algoritmo genético de chaves aleatórias viciadas para a solução do problema de reconfiguração de sistemas de distribuição. Esse problema de otimização consiste em encontrar a configuração radial que apresenta perdas mínimas, satisfazendo as restrições topológicas e as restrições operacionais, sendo modelado como um problema de Programação Não Linear Inteira Mista. O método proposto utiliza o algoritmo de Prim na geração de configurações radiais e emprega um algoritmo de fluxo de carga de varredura para avaliar cada proposta de solução. O algoritmo genético de chaves aleatórias viciadas foi desenvolvido na linguagem de programação FORTRAN e foi testado em quatro sistemas de distribuição da literatura especializada (14 barras, 33 barras, 84 barras e 136 barras). Os resultados obtidos da aplicação do algoritmo permitem avaliar o seu desempenho e eficiência em comparação com a melhor solução encontrada na literatura especializada. / Abstract: The application of the biased random-key genetic algorithm for the reconfiguration of distribution systems is proposed in this Dissertation. The problem of reconfiguration in distribution systems consists of finding the radial configuration that presents the minimum losses, satisfying topological and operating constraints and is commonly modeled as a mixed-integer nonlinear programming problem. The proposed method uses the Prim's algorithm to generate radial configurations that are evaluated through a backward/forward sweep power flow method. The biased random-key genetic algorithm used was developed in the programming language FORTRAN and was tested in four systems (14-bus, 33-bus, 84-bus and 136-bus). The obtained results show the performance and efficiency of the proposed method in comparison to the best solution found in the specialized literature. / Mestre

Submitted by Rommel Gregorio Vargas Peralta (rgvp88@gmail.com) on 2018-07-11T22:12:08Z No. of bitstreams: 1 Dissertação Mestrado - Rommel Gregorio Vargas Peralta.pdf: 3543837 bytes, checksum: 233f1b03c9cd16d58a1c981ab331ee0f (MD5) / Approved for entry into archive by Cristina Alexandra de Godoy null (cristina@adm.feis.unesp.br) on 2018-07-12T20:17:35Z (GMT) No. of bitstreams: 1 vargasperalta_rg_me_ilha.pdf: 3732403 bytes, checksum: eb45069beb69cdfd8b2cc75bf47668bf (MD5) / Made available in DSpace on 2018-07-12T20:17:35Z (GMT). No. of bitstreams: 1 vargasperalta_rg_me_ilha.pdf: 3732403 bytes, checksum: eb45069beb69cdfd8b2cc75bf47668bf (MD5) Previous issue date: 2018-04-20 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Nesta dissertação é proposta a aplicação do algoritmo genético de chaves aleatórias viciadas para a solução do problema de reconfiguração de sistemas de distribuição. Esse problema de otimização consiste em encontrar a configuração radial que apresenta perdas mínimas, satisfazendo as restrições topológicas e as restrições operacionais, sendo modelado como um problema de Programação Não Linear Inteira Mista. O método proposto utiliza o algoritmo de Prim na geração de configurações radiais e emprega um algoritmo de fluxo de carga de varredura para avaliar cada proposta de solução. O algoritmo genético de chaves aleatórias viciadas foi desenvolvido na linguagem de programação FORTRAN e foi testado em quatro sistemas de distribuição da literatura especializada (14 barras, 33 barras, 84 barras e 136 barras). Os resultados obtidos da aplicação do algoritmo permitem avaliar o seu desempenho e eficiência em comparação com a melhor solução encontrada na literatura especializada. / The application of the biased random-key genetic algorithm for the reconfiguration of distribution systems is proposed in this Dissertation. The problem of reconfiguration in distribution systems consists of finding the radial configuration that presents the minimum losses, satisfying topological and operating constraints and is commonly modeled as a mixed-integer nonlinear programming problem. The proposed method uses the Prim's algorithm to generate radial configurations that are evaluated through a backward/forward sweep power flow method. The biased random-key genetic algorithm used was developed in the programming language FORTRAN and was tested in four systems (14-bus, 33-bus, 84-bus and 136-bus). The obtained results show the performance and efficiency of the proposed method in comparison to the best solution found in the specialized literature.

Informações sobre as características funcionais e estruturais da unidade motora (UM) são altamente relevantes em investigações fisiológicas e nos estudos clínicos das disfunções neuromusculares. A eletromiografia (EMG) é um método adequado para obtenção dessas informações. Entretanto, devido à dificuldade na separação da atividade individual de uma unidade motora das outras que estão simultaneamente ativas, seu uso em clínica prática se dá comumente através de métodos invasivos, empregando eletrodos de agulha ou fios implantados. Apesar da EMG de superfície ser não-invasiva e, portanto mais apropriada para aplicações clínicas, não é usada em clínica porque não há até o presente um método satisfatório para decomposição do sinal EMG de superfície. Um EMG de superfície é muito mais difícil de decompor devido a significante superposição dos Potenciais de Ação das UMs (MUAPs) e a relação sinal-ruído relativamente baixa, se comparada aos métodos invasivos. Defendemos que a separação da atividade individual das UMs pode ser feita de modo não-invasivo aliando-se técnicas de aquisição altamente especializadas com técnicas usadas em reconhecimento de padrões. Desenvolvemos um método para decomposição de EMGs de superfície, a partir do qual foi possível extrair características relevantes das UMs, que permitem seu uso em avaliação e diagnóstico de desordens neuromusculares. Em nossa abordagem, o sinal EMG é inicialmente captado sob contração isométrica fraca usando eletrodos desuperfície. O sinal EMG bruto passa em seguida por um filtro Diferencial Passa-Baixas Ponderado (DPBP) em série com um detector de picos, que detecta os picos de MUAPs e extrai suas formas de onda. Na sequência, o conjunto de MUAPs extraído é classificado por uma rede neural SOM, e os MUAPs agrupados pela similaridade de suas formas de onda. No próximo passo a informação temporal dos disparos é checada, eliminando possíveis erros de classificação, e finalmente os Trens de MUAPs (MUAPTs) das UMs individuais são reconstituídos do EMG original. As estatísticas de disparos (IPI) bem como as formas de ondas dos MUAPs das respectivas UMs são então extraídas e armazenadas para estudos posteriores. Resultados preliminares obtidos com EMGs normais e patológicos, extraídos de membros superiores sob contração fraca, indicam que, o método mostrou-se apto a decompor EMGs de superfícies, além de potencial para aplicações em estudos clínicos não-invasivos de disfunções neuromusculares.Informações sobre as características funcionais e estruturais da unidade motora (UM) são altamente relevantes em investigações fisiológicas e nos estudos clínicos das disfunções neuromusculares. A eletromiografia (EMG) é um método adequado para obtenção dessas informações. Entretanto, devido à dificuldade na separação da atividade individual de uma unidade motora das outras que estão simultaneamente ativas, seu uso em clínica prática se dá comumente através de métodos invasivos, empregando eletrodos de agulha ou fios implantados. Apesar da EMG de superfície ser não-invasiva e, portanto mais apropriada para aplicações clínicas, não é usada em clínica porque não há até o presente um método satisfatório para decomposição do sinal EMG de superfície. Um EMG de superfície é muito mais difícil de decompor devido a significante superposição dos Potenciais de Ação das UMs (MUAPs) e a relação sinal-ruído relativamente baixa, se comparada aos métodos invasivos. Defendemos que a separação da atividade individual das UMs pode ser feita de modo não-invasivo aliando-se técnicas de aquisição altamente especializadas com técnicas usadas em reconhecimento de padrões. Desenvolvemos um método para decomposição de EMGs de superfície, a partir do qual foi possível extrair características relevantes das UMs, que permitem seu uso em avaliação e diagnóstico de desordens neuromusculares. Em nossa abordagem, o sinal EMG é inicialmente captado sob contração isométrica fraca usando eletrodos desuperfície. O sinal EMG bruto passa em seguida por um filtro Diferencial Passa-Baixas Ponderado (DPBP) em série com um detector de picos, que detecta os picos de MUAPs e extrai suas formas de onda. Na sequência, o conjunto de MUAPs extraído é classificado por uma rede neural SOM, e os MUAPs agrupados pela similaridade de suas formas de onda. No próximo passo a informação temporal dos disparos é checada, eliminando possíveis erros de classificação, e finalmente os Trens de MUAPs (MUAPTs) das UMs individuais são reconstituídos do EMG original. As estatísticas de disparos (IPI) bem como as formas de ondas dos MUAPs das respectivas UMs são então extraídas e armazenadas para estudos posteriores. Resultados preliminares obtidos com EMGs normais e patológicos, extraídos de membros superiores sob contração fraca, indicam que, o método mostrou-se apto a decompor EMGs de superfícies, além de potencial para aplicações em estudos clínicos não-invasivos de disfunções neuromusculares. / Information on the functional and structural characteristics of the motor unit (MU) they are highly important in physiologic investigations and in the clinical studies of the neuromuscular dysfunctions. The electromyography (EMG) it is an appropriate method for obtaining of that information. However, due to the difficulty in the separation of the individual activity of a motor unit of the another that are simultaneously active, your use in practical clinic happen commonly through methods invasive, employing needle electrodes or implanted threads. In spite of surface EMG to be non-invasive and, therefore more appropriate for clinical applications, it is not used at clinic because there is not until the present a satisfactory method for decomposition of the surface EMG sign. A surface EMG is much more difficult of decomposing due to significant overlap of the Motor Unit Action Potentials (MUAPs) and the relationship sign-noise relatively low, if compared to the invasive methods. We defended that the separation of the individual activity of MUs can be made in way non-invasive allying highly specialized acquisition techniques with techniques used in recognition of patterns. We developed a method for decomposition of surface EMGs, starting from which was possible to extract important characteristics of MUs, which allow your use in evaluation and diagnosis of neuromuscular disorders. In our approach, the sign EMG is captured initially under weak isometriccontraction using surface electrodes. The sign EMG raw raisin soon after for a Biased Low-Pass Differential filter (BLPD) in series with a detector of peaks, that detects the peaks of MUAPs and it extracts your wave forms. In the sequence, a SOM neural network classifies the set of extracted MUAPs, and MUAPs are clustered by the similarity in your wave shape. In the next step the temporal information of the discharges is checked, eliminating possible classification mistakes, and finally the MUAPs Trains (MUAPTs) of individual MUs they are reconstituted of original EMG. The statistics of discharges (IPI) as well as the forms of waves of MUAPs of respective MUs are then extracted and stored for subsequent studies. Results preliminaries obtained with normal and pathological EMGs, extracted of superior members under weak contraction, they indicate that, the method was shown capable to decompose surfaces EMGs, besides potential for applications in clinical studies non-invasive of neuromuscular dysfunctions.

Malgré leur importance dans le déterminisme du sexe chez de nombreux organismes, les chromosomes sexuels ont été étudiés chez quelques espèces seulement du fait du manque de séquences disponibles. En effet, le séquençage et l'assemblage des chromosomes sexuels est rendu très difficile par leurs abondantes séquences répétées. Durant cette thèse, une méthode probabiliste a été développée pour inférer les gènes liés au sexe à partir de données RNA-seq chez une famille. Des tests de cette méthode appelée SEX-DETector sur des données réelles et simulées suggèrent qu'elle fonctionnera sur une grande variété de systèmes. La méthode a inféré ∼1300 gènes liés au sexe chez Silene latifolia, une plante dioïque qui possède des chromosomes sexuels XY pour lesquels quelques données de séquence sont disponibles (dont certaines obtenues lors de cette thèse par séquençage de BACs). Les gènes du Y sont moins exprimés que ceux du X chez S. latifolia, mais le statut de la compensation de dosage (un mécanisme qui corrige la sous-expression des gènes liés au sexe chez les males) est encore controversé. L'analyse des nouveaux gènes liés au sexe inférés par SEX-DETector a permis de confirmer la compensation de dosage chez S. latifolia, qui est effectuée par la surexpression du X maternel, possiblement via un mécanisme epigénétique d'empreinte. Les données ont également été utilisées pour étudier l'évolution de l'expression biaisée pour le sexe chez S. latifolia et ont révélé que la majorité des changements de niveaux d'expression ont eu lieu chez les femelles. Les implications de nos résultats concernant l'évolution de la dioécie et des chromosomes sexuels sont discutés / In many organisms, sexes are determined by sex chromosomes. However, studies have been greatly limited by the paucity of sex chromosome sequences. Indeed, sequencing and assembling sex chromosomes are very challenging due to the large quantity of repetitive DNA that these chromosomes comprise. In this PhD, a probabilistic method was developed to infer sex-linked genes from RNA-seq data of a family (parents and progeny of each sex). The method, called SEX-DETector, was tested on simulated and real data and should performwell on a wide variety of sex chomosome systems. This new method was applied to Silene latifolia, a dioecious plant with XY system, for which partial sequence data on sex chromosomes are available (some of which obtained during this PhD by BAC sequencing), SEX-DETector returned ∼1300 sex-linked genes. In S. latifolia, Y genes are less expressed than their X counterparts. Dosage compensation (a mechanism that corrects for reduced dosage due to Y degeneration in males) was previously tested in S. latifolia, but different studies returned conflicting results. The analysis of the new set of sex-linked genes confirmed the existence of dosage compensation in S. latifolia, which seems to be achieved by the hyperexpression of the maternal X chromosome in males. An imprinting mechanism might underlie dosage compensation in that species. The RNAseq datawere also used to study the evolution of differential expression among sexes in S. latifolia, and revealed that in this species most changes have affected the female sex. The implications of our results for the evolution of dioecy and sex chromosomes in plants are discussed

Le paysage nucléotidique – la manière dont la composition nucléotidique varie le long du génome – est une caractéristique marquante de l'organisation des génomes et varie fortement entre espèces. Plusieurs hypothèses émergent des nombreux débats autour des mécanismes évolutifs à l'origine de ces hétérogénéités du taux de GC, parmi lesquelles la conversion génique biaisée vers G et C (BGC) et la sélection sur l'usage du code (SUC). La BGC est un processus neutre associé à la recombinaison qui favorise les allèles en G ou C au détriment des allèles en A ou T. La SUC est une force de sélection qui favorise les codons dits « préférés », ceux dont la traduction serait la plus efficace. Contrairement à ceux des vertébrés, les paysages nucléotidiques des plantes sont peu connus. La plupart des études ont été consacrées au génome d'Arabidopsis thaliana, pauvre en GC et homogène, et à celui du riz, riche en GC et hétérogène. Le contraste entre ces deux génomes a souvent été généralisé comme une dichotomie entre dicotylédones et monocotylédones, mais cette vision est clairement phylogénétiquement biaisée.Les objectifs de ce travail de thèse sont de caractériser les paysages nucléotidiques des angiospermes à une large échelle phylogénétique et de mieux comprendre quels sont les mécanismes évolutifs jouant sur l'évolution de ces paysages nucléotidiques. Comment varient les paysages nucléotidiques le long de la phylogénie des angiospermes ? SUC et BGC façonnent-elles ces paysages nucléotidiques ? Les différents taxons sont-ils affectés avec la même intensité ?Pour répondre à ces questions, j'ai utilisé une approche de génomique comparative basée sur l'analyse de données EST chez plus de 230 espèces d'angiospermes et de gymnospermes. L'exploration des paysages nucléotidiques de ce large éventail de plantes a montré que les patrons d'hétérogénéité des paysages nucléotidiques suivent un continuum le long de la phylogénie, avec des groupes particulièrement riches et hétérogènes en GC, les graminées par exemple. Mes résultats suggèrent que les paysages nucléotidiques des plantes pourraient avoir été façonnés par la BGC et, dans une moindre mesure, par la SUC. Des épisodes indépendants d'enrichissement et d'appauvrissement en GC ont vraisemblablement eu lieu au cours de l'évolution des plantes, et pourraient être expliqués par des variations d'intensité de ces mécanismes. En utilisant une prédiction du degré d'expression des EST, j'ai également mis en évidence une diversité dans les codons préférés entre espèces. Les préférences d'usage des codons se sont révélées plus labiles au cours de l'évolution pour les codons des acides aminés au code quatre et six fois dégénéré. J'ai pu lier l'évolution des préférences d'usage des codons à l'évolution de la composition nucléotidique des génomes. Mes résultats suggèrent que la composition en base des génomes, affectée en partie par la BGC, orienterait la coévolution entre préférence d'usage du code et ARNt. / The nucleotide landscape – the way base composition varies along a genome – is a striking feature of genome organization and is highly variable between species. The evolutionary causes of such heterogeneity in GC content have been much debated. Biased gene conversion towards G and C (BGC) and selection on codon usage (SCU) are thought to be main forces. BGC is a neutral process associated with recombination favouring G and C alleles over A and T ones. SCU is a selection process favouring the so-called “preferred” codons, i.e., those whose translation is the most efficient. Contrary to vertebrates, plant nucleotide landscapes are still poorly known. Most studies focused on the GC-poor and homogeneous Arabidopsis thaliana genome and on the GC-rich and heterogeneous rice genome. The contrast between these two genomes was often generalized as a dicot/monocot dichotomy but this vision is clearly phylogenetically biased.The objectives of this study are to characterize angiosperm nucleotide landscapes on a wide phylogenetic scale and to better understand the evolutionary mechanisms acting upon the evolution of nucleotide landscapes. To what extent do nucleotide landscapes vary across angiosperm phylogeny? Are nucleotide landscapes shaped by BGC and SCU? Are taxa affected with the same intensity?To tackle these issues, I used a comparative genomic approach relying on EST data analysis on over 230 angiosperm and gymnosperm species. Through the nucleotide landscape survey for such a wide range of species I found a continuum of GC-heterogeneity patterns across phylogeny, some taxa such as Poaceae being strikingly GC-rich and heterogeneous. My results suggest that nucleotide landscapes could have been shaped by BGC and, to a lesser extent, by SCU. GC-content enrichment and impoverishment are likely to have occurred several times independently during plant evolution and could be explained by intensity variations of BGC and SCU. Using a proxy for EST expression level, I also characterized the diversity of preferred codons between species. Codon usage preferences were shown to be evolutionarily more unstable for four- and six-fold degenerate codon families. Finally, I could link the evolution of codon usage preferences to the evolution of genome base composition. My results suggest that genome base composition, partially shaped by BGC, seems to drive the coevolution between codon usage preferences and tRNAs.