La réalisation de microsystèmes de multidétection et sans marquage pour la reconnaissance de biomolécules est d'un intérêt fondamental pour la réalisation de tests rapides dédiés au diagnostic biologique. Ces développements nécessitent une méthode d'adressage des sondes de capture sur une plateforme multiplexée associée à une méthode d'analyse sensible. Dans cette étude, la méthode de détection choisie pour les tests développés sur la puce est la voltampérométrie cyclique, et le férrocène a été utilisé pour la modification de sondes oligonucléotidiques de type tige-boucle. Une stratégie d'électroadressage a été développée sur surface d'or. Elle a été réalisée via chimie « click » entre un alcyne et un azoture. Cette réaction peut être électro-catalysée en maintenant le catalyseur cuivre sous sa forme active par l'application d'un potentiel à l'électrode. Une première entité chimique de petite taille, constituée de deux groupements dithiol phosphate et d'un groupement hexynyle a été synthétisé par synthèse supportée et greffée sur électrode d'or. Par la suite, différents éléments ont été immobilisés par chimie « click ». Un dérivé ferrocène porteur d'une fonction azoture a été utilisé pour la détermination des conditions optimales de cette chimie. Puis, cette méthode a été exploitée pour l'immobilisation de nanoparticules fluorescentes et de protéines par l'intermédiaire de la formation du complexe biotine/streptavidine. Enfin, cette méthode a permis l'électroadressage de sondes de capture oligonucléotidiques de type tige-boucle, modifiées par des ferrocènes. Des tests d'hybridation ADN ont été menés en milieu complexe avec une limite de détection déterminée à 100fM / This production of microsystem for label-free multi detection of biomolecules is fundamental for the realization of rapid tests dedicated to laboratory diagnosis. A viable method is requires to both address capture probes and to be associates with a sensitive analysis on multiplexed platform. In this study, the method chosen for detection on electrode is cyclic voltammetry, and ferrocene was used to modify stem-loop oligonucleotides. A strategy was developed for the electroadressing of probes on gold surface. It is performed through chemistry “click” between an alkyne and an azide. This reaction may be catalyzed by maintaining the correct potential to the electrode to form an active copper oxidation state on the surface. A first small chemical entity, containing two phosphate dithiol moieties and a hexynyl moiety was synthesized by supported chemistry and grafted on gold electrode. Thereafter, various elements were immobilized by chemistry “click”. Ferrocene derivative carrying an azide function was used to determine the optimal conditions for this chemistry. Then, this method has been exploited for the immobilization of proteins and fluorescent nanoparticles via the formation of biotin/streptavidin complex. Finally, this method allowed to electroaddress stem-loop oligonucleotids, designed as capture probes, modified by ferrocene. DNA hybridization tests were conducted in complex environments with a detection limit determined at 100 fM

L'endocardite infectieuse (EI) est une complication rare mais gravissime de la bactériémie à Staphylococcus aureus. Bien que certains facteurs de risque liés à l'hôte aient été décrits, l'implication de facteurs bactériens dans la survenue de l'EI est encore inconnue. Ces travaux de thèse ont visé à chercher tout élément bactérien associé à l'EI. Les facteurs phénotypiques décrits ou supposés comme potentiellement impliqués dans l'EI ont été testés. En parallèle, les profils génotypiques des souches obtenus par puces ADN ont été analysés par différents outils statistiques. L'analyse statistique univariée n'a montré aucune différence significative entre souches d'EI et souches de bactériémie, suggérant un processus complexe et multifactoriel. En effet, l'analyse discriminante en composante principale appliquée sur les données de puces ADN a permis de mettre en évidence une distinction entre les deux groupes de souches, confirmée sur une collection indépendante de souches. De plus, une fonction linéaire simplifiée, basée sur seulement 8 marqueurs génétiques, a permis d'obtenir des performances similaires, sur la collection de souches initiale ainsi que la collection indépendante de validation. En dernier lieu, les souches d'EI et de bactériémie ont été comparées à partir de séquences du génome complet (n = 40 (20 EI, 20 bactériémies)). L'analyse statistique par analyse discriminante en composante principale réalisée sur ces données génomiques confirme une distinction possible entre les deux groupes de souches. Au total, ces travaux de thèse apportent la preuve de concept que les facteurs bactériens sont impliqués dans la survenue de l'EI au cours de bactériémie à S. aureus / Infective endocarditis (IE) is a severe condition complicating 10-25% of Staphylococcus aureus bacteremia. Although host-related IE risk factors have been identified, the involvement of bacterial features in IE complication is still unclear. This PhD work aimed to characterize strictly defined IE and bacteremia isolates and searched for discriminant features. Phenotypic traits previously reported or hypothesized to be involved in staphylococcal IE pathogenesis were tested. In parallel, the genotypic profiles of all isolates, obtained by microarray, were analyzed. No significant difference was observed between IE and bacteremia strains, regarding either phenotypic or genotypic univariate analyses, suggesting a multifactorial process. However, the discriminant analysis of principal components (DAPC), applied on microarray data, segregated IE and bacteremia isolates. The performance of this model was confirmed with an independent collection of IE and bacteremia isolates. Finally, a simple linear discriminant function based on a subset of 8 genetic markers retained valuable performance both in study collection and in the independent validation collection. At last, IE and bacteremia isolates were compared based on whole genome sequence data from a subset of 40 isolates. When applied to this dataset, DAPC confirmed a possible segregation between the two groups of isolates. All in all, this PhD work provides the proof of concept that bacterial characteristics may contribute to the occurrence of IE in patients with S. aureus bacteremia

O câncer de próstata é o quinto tipo mais comum de câncer no mundo e o mais comum em homens. Fatores clínicos e anatomopatológicos atualmente usados na clínica não são capazes de distinguir entre a doença indolente e a agressiva. Existe uma grande necessidade de novos marcadores de prognóstico, a fim de melhorar o gerenciamento clínico de pacientes de câncer de próstata. Além das anormalidades em genes codificadores de proteínas, alterações em RNAs não codificadores (ncRNAs) contribuem para a patogênese do câncer e, portanto, representam outra fonte potencial de biomarcadores de câncer de próstata. Entretanto, até o momento, poucos estudos de perfis de expressão de ncRNAs foram publicados. Este projeto teve como principal objetivo identificar perfis de expressão de genes codificadores e não codificadores de proteína correlacionados com recorrência de tumor de próstata, a fim de gerar um perfil prognóstico com potencial uso como biomarcadores e elucidar o possível papel de ncRNAs no desenvolvimento do câncer. Para isso, foram analisados os perfis de expressão de genes codificadores e não codificadores de proteína de um conjunto de 42 amostras de tecido tumoral de câncer de próstata de pacientes de amostras de pacientes submetidos à prostatectomia radical, com longo acompanhamento clínico (cinco anos) e conhecida evolução da doença Nós utilizamos microarranjos por nós desenhados e fabricados pela Agilent sob encomenda, interrogando aproximadamente 18.709 transcritos não codificadores longos (>500 nt), sem evidência de splicing, que mapeiam em regiões intrônicas dentro de 5.660 loci genômicos. Os dados de expressão foram extraídos de cada arranjo, normalizados entre todas as 42 amostras de pacientes. Usando uma estratégia de múltipla amostragem, foi identificado um perfil de expressão de mau prognóstico, contendo 51 transcritos intrônicos não codificadores de proteína. O perfil prognóstico de ncRNAs foi aplicado a um conjunto teste independente de 22 pacientes, classificando corretamente 82% das amostras. Uma análise de Kaplan-Meier dos pacientes do conjunto teste indicou que as curvas de sobrevida dos grupos de alto e baixo risco foram significativamente distintas (Log-rank test p = 0,0009; Hazard ratio = 23,4, 95% CI = 3,62 a 151,2), confirmando assim que este classificador é útil para identificar pacientes com alto risco de recorrência. Além disso, estas descobertas indicam um potencial papel destes RNAs intrônicos não codificadores na progressão do tumor de próstata e apontam para os RNAs intrônicos como potenciais novos marcadores de câncer / Prostate cancer is the fifth most common type of cancer in the world, and the most common in men. Clinical and anatomo-pathological factors currently used in clinic are not able to distinguish between the indolent and the aggressive disease. There is a major need of new prognostic makers in order to improve the clinical management of prostate cancer patients. Apart from abnormalities in protein-coding genes, changes in non-coding RNAs (ncRNAs) contribute to the pathogenesis of cancer and thus represent another potential source of prostate cancer biomarkers. However, few studies of expression profiles of ncRNAs have been published. This project aimed to identify expression profiles of protein-coding and non-coding genes correlated to prostate cancer biochemical recurrence. For this, we analyzed the expression profile of 42 prostate cancer samples from patients undergoing radical prostatectomy, with long follow-up (five years), and know disease outcome. We used a custom microarray designed by us and printed by Agilent, that probes 18,709 long (>500 nt) ncRNAs mapping to intronic regions within 5,660 genomic loci. The expression data were extracted from each array and normalized across all 42 samples. Using a multiple random sampling validation strategy, we identified an expression profile of poor prognosis, comprising 51 ncRNAs. The prognostic profile of ncRNAs was applied to an independent test set of 22 patients, correctly classifying 82% of the samples. A Kaplan-Meier analysis of the test set of patients indicated that the survival curves of high and low risk groups were significantly different (Log-rank test p = 0.0009, Hazard ratio = 23.4, 95% CI = 3.62 to 151.2) thus confirming that this classifier is useful for identifying patients at high risk of recurrence. Furthermore, these findings indicate a potential role of these intronic non-coding RNAs in the progression of prostate tumors and points to the intronic ncRNAs as potential new markers of cancer.

L’ARN polymérase dépendante d’ADN (RNAP) est une enzyme responsable de la polymérisation de ribonucleotides dans une séquence d'ARN complémentaire de l'ADN de matrice. La famille de RNAP a plusieurs membres, comme de protéines sous-unité unique (par exemple du bactériophage T7) ou multiple sous-unité (bactériennes et eucaryotes). Transcription de l'ARN - un événement crucial dans l'expression des gènes - varie en fonction de l'origine de RNAP. Bien que le processus de transcription est relativement bien caractérisée, de nombreux éléments restent mal compris, surtout par rapport à la dynamique de la reconnaissance de promoteur, d'évasion et de l'allongement dans une contexte de cellule où la densité moléculaire, les concentrations et les effets plus proches environs sont importants. L'objectif de cette thèse était la développement d’une méthode qui permettrait suivre la réaction RNAP in vitro en temps réel dans des conditions très contrôlées. Un axe majeur a été mis pour développer un biocapteur basé surface qui permettrait à la caractérisation des principales étapes de la réaction de transcription. Par conséquent, les interactions entre des molécules d'ADN immobilisés sur une surface du capteur et RNAP libre délivré par un système microfluidique de la surface ont été examinées. Changements de l'indice de réfraction, corrélés avec les changements de masse sur la surface ont été suivis en utilisant l'imagerie par résonance de plasmon de surface (SPRi). SPRi est une technique sensible dédiée à l'analyse des interactions entre deux ligands en temps réel. Les bases du mécanisme sont la détection de légères différences dans la réflexion de la lumière polarisée à un angle fixe qui est associé avec une variation de masse à l'interface. Les données obtenues à partir SPRi sont utilisées pour déterminer la cinétique des interactions. Géométrie d’ADN puces permet de suivre plusieurs échantillons simultanément, qui raccourcit considérablement le temps que de manipulation et améliore la qualité et la reproductibilité des résultats obtenus. Autres biocapteurs optofluidique: résonateur de microring et microscopie de fluorescence par réflexion totale interne (TIRF) ont été développés en parallèle. Nous avons biofunctionalisé et caractérisé des surfaces de capteur (de verre couvert de polymère pour un résonateur de microring et la microscopie TIRF et 50 nm couche mince d’or sur des prismes de SPRi) afin d'immobiliser ADN d'une manière contrôlée, par création d’une monocouche auto-assemblée (SAM). Fonctionnalisation de polymères SU-8 concernées deux méthodes: covalent immobilisation de (bio) molécules et la conjugaison non covalente sur la base de couplage hydrophobe. Pour la fonctionnalisation de surface d’or, quatre stratégies différentes d'immobilisation des molécules ont été comparés: formation de la liaison de thiol - or, la formation des liaisons amide, interactions extrAvidin - biotine et le couplage hydrophobe. Les études de la conjugaison de l'ADN à la surface d'or fonctionnalisé ont été effectuées en ce qui concerne la spécificité et la densité d'ADN immobilisées de longueurs différentes: 50, 500 et 1000 pb. Enfin, les surfaces biofunctionalized ont été utilisées pour suivre en temps réel des réactions de transcription de deux RNAP: bactériophage T7 RNAP monomère et l'holoenzyme d'Escherichia coli RNAP. Les analyses cinétiques de la formation d’un complexe nucléoprotéine et la transcription d'ARN ont été fait par report de la densité et la longueur de l'ADN immobilisé, la position de la séquence du promoteur spécifique. Transcription de l'ARN dans l'appareil SPRi a été confirmée par la collection, la détection et l'analyse des produits ARN.L'objectif final comprends une synthèse de l'ARN contrôlée qui serait une étape intermédiaire d'enquêter en temps réel la production de protéines in vitro. / DNA-dependent RNA polymerase (RNAP) is an enzyme responsible for the polymerization of ribonucleotides into an RNA sequence complementary to the template DNA. RNAP family has several members being single subunit (e.g. T7 bacteriophage) or multi subunit (bacterial and eukaryote) proteins. RNA transcription – a crucial event in gene expression – differs depending on the RNAP origin. Although the transcription process is relatively well characterized, many elements remain poorly understood, especially with respect to the dynamics of promoter recognition, escape and elongation in a cell like context where molecular density, concentrations and nearest neighbour effects are prevalent.The goal of this thesis was to develop a robust method that would allow real time monitoring of RNAP reaction in vitro in thoroughly controlled conditions. A major axis was to develop a surface-based biosensor that would allow the characterization of the main steps of the transcription reaction. Consequently, interactions between DNA molecules immobilized on a sensor surface and free RNAP delivered through a microfluidic flow system to the surface were examined. Changes in refractive index, correlated with changes in mass at a surface were followed using surface plasmon resonance imaging (SPRi). SPRi is a sensitive technique dedicated to analysis of interactions between two ligands in real time. The mechanism bases on the detection of slight differences in the reflectivity of polarized light at a fixed angle that are associated with a mass variation at the interface. Data obtained from SPRi are used to determine the kinetics of the interactions. Microarray geometry of SPRi allows monitoring several samples simultaneously that significantly shortens manipulation time and improves a quality and reproducibility of obtained results. Other label-free optofluidic biosensors: microring resonator and total internal reflection fluorescence (TIRF) microscopy were developed in parallel.We firstly biofunctionalized and characterized sensor surfaces (polymer coated glass for microring resonator and TIRF microscopy and 50-nm thin layer gold coatings on glass prisms for SPRi) in order to immobilize DNA strands in a controlled manner, using a self-assembled monolayer (SAM). Functionalization of photoresist polymer SU-8 concerned two methods: covalent (bio)molecule grafting and non-covalent conjugation based on hydrophobic coupling. Regarding gold surface functionalization, four different strategies of antifouling (bio)molecule immobilization were compared: thiol – gold bond formation, amide bond formation, extrAvidin – biotin interactions and hydrophobic coupling. Studies of DNA conjugation to the functionalized gold surface were performed with respect to specificity and density of immobilized DNA molecules of different lengths: 50, 500 and 1000 bp.Finally, biofunctionalized surfaces were used for real time monitoring of transcription reactions using two RNAPs: monomeric bacteriophage T7 RNAP and the holoenzyme of Escherichia coli RNAP. Kinetic analyses of nucleoprotein complex formation and RNA transcription were performed as a function of immobilized DNA density, the length of the immobilized DNA, the position of the specific promoter sequence with respect to the point of immobilization and the direction of subsequent transcription. RNA transcription in the SPRi apparatus was confirmed by collection, detection and analysis of relevant products.The future development of biosensors dedicated to in vitro gene expression will include the adaptation of the methods presented above to other optofluidic systems and further development of the technique. The final goal comprises a controlled RNA synthesis that would be an intermediate step to investigate real time in vitro protein production.

No description available.

Les microorganismes constituent la plus grande diversité du monde vivant. Ils jouent un rôle clef dans tous les processus biologiques grâce à leurs capacités d’adaptation et à la diversité de leurs capacités métaboliques. Le développement de nouvelles approches de génomique permet de mieux explorer les populations microbiennes. Dans ce contexte, les biopuces à ADN représentent un outil à haut débit de choix pour l'étude de plusieurs milliers d’espèces en une seule expérience. Cependant, la conception et l’analyse des biopuces à ADN, avec leurs formats de haute densité actuels ainsi que l’immense quantité de données à traiter, représentent des étapes complexes mais cruciales. Pour améliorer la qualité et la performance de ces deux étapes, nous avons proposé de nouvelles approches bioinformatiques pour la conception et l’analyse des biopuces à ADN en environnements parallèles. Ces approches généralistes et polyvalentes utilisent le calcul haute performance (HPC) et les nouvelles approches du génie logiciel inspirées de la modélisation, notamment l’ingénierie dirigée par les modèles (IDM) pour contourner les limites actuelles. Nous avons développé PhylGrid 2.0, une nouvelle approche distribuée sur grilles de calcul pour la sélection de sondes exploratoires pour biopuces phylogénétiques. Ce logiciel a alors été utilisé pour construire PhylOPDb: une base de données complète de sondes oligonucléotidiques pour l’étude des communautés procaryotiques. MetaExploArrays qui est un logiciel parallèle pour la détermination de sondes sur différentes architectures de calcul (un PC, un multiprocesseur, un cluster ou une grille de calcul), en utilisant une approche de méta-programmation et d’ingénierie dirigée par les modèles a alors été conçu pour apporter une flexibilité aux utilisateurs en fonction de leurs ressources matériel. PhylInterpret, quant à lui est un nouveau logiciel pour faciliter l’analyse des résultats d’hybridation des biopuces à ADN. PhylInterpret utilise les notions de la logique propositionnelle pour déterminer la composition en procaryotes d’échantillons métagénomiques. Enfin, une démarche d’ingénierie dirigée par les modèles pour la parallélisation de la traduction inverse d’oligopeptides pour le design des biopuces à ADN fonctionnelles a également été mise en place. / Microorganisms represent the largest diversity of the living beings. They play a crucial rôle in all biological processes related to their huge metabolic potentialities and their capacity for adaptation to different ecological niches. The development of new genomic approaches allows a better knowledge of the microbial communities involved in complex environments functioning. In this context, DNA microarrays represent high-throughput tools able to study the presence, or the expression levels of several thousands of genes, combining qualitative and quantitative aspects in only one experiment. However, the design and analysis of DNA microarrays, with their current high density formats as well as the huge amount of data to process, are complex but crucial steps. To improve the quality and performance of these two steps, we have proposed new bioinformatics approaches for the design and analysis of DNA microarrays in parallel and distributed environments. These multipurpose approaches use high performance computing (HPC) and new software engineering approaches, especially model driven engineering (MDE), to overcome the current limitations. We have first developed PhylGrid 2.0, a new distributed approach for the selection of explorative probes for phylogenetic DNA microarrays at large scale using computing grids. This software was used to build PhylOPDb: a comprehensive 16S rRNA oligonucleotide probe database for prokaryotic identification. MetaExploArrays, which is a parallel software of oligonucleotide probe selection on different computing architectures (a PC, a multiprocessor, a cluster or a computing grid) using meta-programming and a model driven engineering approach, has been developed to improve flexibility in accordance to user’s informatics resources. Then, PhylInterpret, a new software for the analysis of hybridization results of DNA microarrays. PhylInterpret uses the concepts of propositional logic to determine the prokaryotic composition of metagenomic samples. Finally, a new parallelization method based on model driven engineering (MDE) has been proposed to compute a complete backtranslation of short peptides to select probes for functional microarrays.

A linhagem Pedra 2 (PE-2) de Saccharomyces cerevisiae destaca-se por ser o organismo mais comumente utilizado no processo industrial de produção de biocombustíveis. Com a descoberta das linhagens selvagens, alcançou-se uma maior tolerância ao etanol permitindo o desenvolvimento de uma nova tecnologia: a fermentação com alto teor alcoólico. Desta forma, foi possível aumentar tanto a concentração de açúcares totais do mosto quanto o volume final de etanol purificado. No entanto, esse novo processo de fermentação tem causado um agravamento nos diversos estresses aplicados à levedura. No presente trabalho, é apresentado o perfil transcricional da linhagem PE-2 sob condições de alto etanol em fermentadores industriais através da tecnologia de microarranjo de DNA. Com a utilização desta, analisou-se o perfil global da expressão da levedura, identificando grupos de genes de interesse e vias metabólicas correguladas no processo de adaptação e sobrevivência às diferentes condições de estresses impostas a levedura pela fermentação industrial. Mais especificamente, 5860 genes foram estudados nesse trabalho e tiveram as suas variações de expressão quantificadas ao longo dos tempos 0, 6, 12 e 18 horas do ciclo fermentativo industrial. Em particular, algumas vias metabólicas associadas a compostos-chave no processo fermentativo tiveram seus genes diferenciamente expressos mapeados. Além disso, identificou-se vários grupos de genes altamente correlacionados a diferentes processos biológicos em S. cerevisiae, como por exemplo, a atividade de biossíntese de etanol. Por fim, espera-se que estes resultados forneçam bases para a realização de estudos mais direcionados no intuito de obter uma maior eficiência fermentativa e adaptação a estresses gerados durante durante o processo industrial. / The Pedra-2 (PE-2) strain of Saccharomyces cerevisiae is commonly used in the industrial process for biofuel production. In studies of wild type strains of S. cerevisiae, a wider tolerance to ethanol was achieved, which allowed for the development of a new technology of high alcohol percentage fermentation. This process made possible the increase of total sugar concentration in the mixture, and the volume of purified ethanol, although the new process has caused increase in the stresses applied to the yeast. In this study, the transcriptional profile of the PE-2 strain in high ethanol conditions is presented using DNA microarray. The global expression profile was used to identify groups of genes of interest and to analyze metabolic pathways that became co-regulated in adaptation to stress conditions imposed to the yeast by the industrial fermentation. In particular, 5860 genes were studied in this work and were detailed according to their expression profiles belong the fermentation cicle (0, 6, 12 and 18 hours). Moreover, metabolic pathways associated to key compounds in the fermentative process were described in terms of the composition of the differentially expressed genes. In addition, groups of genes highly correlated to different biological process in S. cerevisiae were identified. Finally, it is expected that this work could provide new directions in the study of fermentative efficiency and induced stress adaptation during the industrial fermentative process.

La dégradation des fibres alimentaires est une fonction essentielle des écosystèmes digestifs microbiens. Chez le ruminant, elle est assurée par des bactéries, champignons et protozoaires capables de produire de nombreuses enzymes nécessaires à l’hydrolyse des polysaccharides de paroi végétale. Parmi les facteurs susceptibles d’influencer l’efficacité de dégradation des fibres, qui est une composante importante de la productivité et de la santé animales, des additifs tels que des levures probiotiques apparaissent comme un levier intéressant. Afin d’approfondir les connaissances sur les facteurs de modulation de l’activité fibrolytique, une biopuce ADN fonctionnelle, outil moléculaire haut-débit, ciblant les gènes codant les enzymes clés de la dégradation de la cellulose et des xylanes dans les écosystèmes digestifs a été développée. Aussi, une méthode efficace dédiée à des échantillons ruminaux pour la soustraction des ARNr à partir des ARN totaux a été mise au point afin d’accroitre la sensibilité de l’outil. La biopuce fonctionnelle a été validée sur échantillons de complexité croissante et démontre d’excellents caractères de spécificité et de sensibilité tout en étant exploratoire et quantitative. Des régulations différentielles de l’arsenal des gènes de la fibrolyse de la bactérie du rumen Fibrobacter succinogenes ont pu être montrées. De même, les résultats sur échantillons de rumen suggèrent un rôle des microorganismes eucaryotes dans la fibrolyse pouvant être plus important qu’initialement envisagé. Cette approche métatranscriptomique dirigée pourra in fine continuer d’être appliquée dans l’étude de l’impact de facteurs biotiques et abiotiques sur la fonction fibrolytique microbienne chez les animaux d’élevage. / Dietary fibre degradation is an essential function of microbial digestive ecosystems. In ruminants, this function is ensured by bacteria, fungi and protozoa, producing a large array of enzymes able to degrade plant cell wall polysaccharides. Among factors likely to influence the efficiency of fibre degradation, which is an important component in animal productivity and health, dietary additives such as probiotic yeasts appear as an interesting tool. To provide more insight on factors modulating fibrolytic activity, we designed a functional DNA microarray targeting genes coding for key enzymes involved in cellulose and xylan degradation by digestive microbiota. Also, an efficient method dedicated to rumen samples for removing microorganisms’ rRNA from total RNA samples was developed to increase the sensitivity of the tool. The DNA microarray was validated using targets of increasing complexity and demonstrated sensitivity and specificity as well as explorative and quantitative potential. Differential expression of genes involved in fibrolysis was evidenced in the rumen bacterium Fibrobacter succinogenes. Moreover, results on rumen samples suggest a more important role of eucaryotes in fibre degradation than previously thought. This targeted metatranscriptomic approach will be further applied to the study of the impact of biotic and abiotic factors on the microbial mechanisms of fibre degradation in livestock.

The generalization abilities of machine learning algorithms often depend on the algorithms’ initialization, parameter settings, training sets, or feature selections. For instance, SVM classifier performance largely relies on whether the selected kernel functions are suitable for real application data. To enhance the performance of individual classifiers, this dissertation proposes classifier fusion models using computational intelligence knowledge to combine different classifiers. The first fusion model called T1FFSVM combines multiple SVM classifiers through constructing a fuzzy logic system. T1FFSVM can be improved by tuning the fuzzy membership functions of linguistic variables using genetic algorithms. The improved model is called GFFSVM. To better handle uncertainties existing in fuzzy MFs and in classification data, T1FFSVM can also be improved by applying type-2 fuzzy logic to construct a type-2 fuzzy classifier fusion model (T2FFSVM). T1FFSVM, GFFSVM, and T2FFSVM use accuracy as a classifier performance measure. AUC (the area under an ROC curve) is proved to be a better classifier performance metric. As a comparison study, AUC-based classifier fusion models are also proposed in the dissertation. The experiments on biomedical datasets demonstrate promising performance of the proposed classifier fusion models comparing with the individual composing classifiers. The proposed classifier fusion models also demonstrate better performance than many existing classifier fusion methods. The dissertation also studies one interesting phenomena in biology domain using machine learning and classifier fusion methods. That is, how protein structures and sequences are related each other. The experiments show that protein segments with similar structures also share similar sequences, which add new insights into the existing knowledge on the relation between protein sequences and structures: similar sequences share high structure similarity, but similar structures may not share high sequence similarity.

A linhagem Pedra 2 (PE-2) de Saccharomyces cerevisiae destaca-se por ser o organismo mais comumente utilizado no processo industrial de produção de biocombustíveis. Com a descoberta das linhagens selvagens, alcançou-se uma maior tolerância ao etanol permitindo o desenvolvimento de uma nova tecnologia: a fermentação com alto teor alcoólico. Desta forma, foi possível aumentar tanto a concentração de açúcares totais do mosto quanto o volume final de etanol purificado. No entanto, esse novo processo de fermentação tem causado um agravamento nos diversos estresses aplicados à levedura. No presente trabalho, é apresentado o perfil transcricional da linhagem PE-2 sob condições de alto etanol em fermentadores industriais através da tecnologia de microarranjo de DNA. Com a utilização desta, analisou-se o perfil global da expressão da levedura, identificando grupos de genes de interesse e vias metabólicas correguladas no processo de adaptação e sobrevivência às diferentes condições de estresses impostas a levedura pela fermentação industrial. Mais especificamente, 5860 genes foram estudados nesse trabalho e tiveram as suas variações de expressão quantificadas ao longo dos tempos 0, 6, 12 e 18 horas do ciclo fermentativo industrial. Em particular, algumas vias metabólicas associadas a compostos-chave no processo fermentativo tiveram seus genes diferenciamente expressos mapeados. Além disso, identificou-se vários grupos de genes altamente correlacionados a diferentes processos biológicos em S. cerevisiae, como por exemplo, a atividade de biossíntese de etanol. Por fim, espera-se que estes resultados forneçam bases para a realização de estudos mais direcionados no intuito de obter uma maior eficiência fermentativa e adaptação a estresses gerados durante durante o processo industrial. / The Pedra-2 (PE-2) strain of Saccharomyces cerevisiae is commonly used in the industrial process for biofuel production. In studies of wild type strains of S. cerevisiae, a wider tolerance to ethanol was achieved, which allowed for the development of a new technology of high alcohol percentage fermentation. This process made possible the increase of total sugar concentration in the mixture, and the volume of purified ethanol, although the new process has caused increase in the stresses applied to the yeast. In this study, the transcriptional profile of the PE-2 strain in high ethanol conditions is presented using DNA microarray. The global expression profile was used to identify groups of genes of interest and to analyze metabolic pathways that became co-regulated in adaptation to stress conditions imposed to the yeast by the industrial fermentation. In particular, 5860 genes were studied in this work and were detailed according to their expression profiles belong the fermentation cicle (0, 6, 12 and 18 hours). Moreover, metabolic pathways associated to key compounds in the fermentative process were described in terms of the composition of the differentially expressed genes. In addition, groups of genes highly correlated to different biological process in S. cerevisiae were identified. Finally, it is expected that this work could provide new directions in the study of fermentative efficiency and induced stress adaptation during the industrial fermentative process.