Philosophiae Doctor - PhD / The estimated 5 × 10³⁰ prokaryotic cells inhabiting our planet sequester some 350–550 Petagrams (1 Pg = 1015 g) of carbon, 85–130 Pg of nitrogen, and 9–14 Pg of phosphorous, making them the largest reservoir of those nutrients on Earth (Whitman et al. 1998). However, reports suggest that only less than 1% of these microscopic organisms are cultivable (Torsvik et al. 1990; Sleator et al. 2008). Until recently with the development of metagenomic techniques, the knowledge of microbial diversity and their metabolic capabilities has been limited to this small fraction of cultivable organisms (Handelsman et al. 1998). While metagenomics has undoubtedly revolutionised the field of microbiology and biotechnology it has been generally acknowledged that the current approaches for metagenomic bio- rospecting / screening have limitations which hinder this approach to fully access the metabolic potentials and genetic variations contained in microbial genomes (Beloqui et al. 2008). In particular, the construction of metagenomic libraries and heterologous expression are amongst the major obstacles. The aim of this study was to develop an ultra-high throughput approach for screening enzyme activities using uncloned metagenomic DNA, thereby eliminating cloning steps, and employing in vitro heterologous expression. To achieve this, three widely used techniques: cell-free transcription-translation, in vitro compartmentalisation (IVC) and Fluorescence Activated Cell Sorting (FACS) were combined to develop this robust technique called metagenomic in vitro compartmentalisation (mIVC-FACS). Moreover, the E. coli commercial cell-free system was used in parallel to a novel, in-house Rhodococcus erythropolis based cell-free system. The versatility of this technique was tested by identifying novel beta-xylosidase encoding genes derived from a thermophilic compost metagenome. In addition, the efficiency of mIVC-FACS was compared to the traditional metagenomic approaches; function-based (clone library screening) and sequence-based (shotgun sequencing and PCR screening). The results obtained here show that the R. erythropolis cell-free system was over thirty-fold more effective than the E. coli based system based on the number of hits obtained per million double emulsions (dE) droplets screened. Six beta-xylosidase encoding genes were isolated and confirmed from twenty-eight positive dE droplets. Most of the droplets that were isolated from the same gate encoded the same enzyme, indicating that this technique is highly selective. A comparison of the hit rate of this screening approach with the traditional E. coli based fosmid library method shows that mIVC-FACS is at least 2.5 times more sensitive. Although only a few hits from the mIVC-FACS screening were selected for confirmation of beta-xylosidase activity, the proposed hit rate suggests that a significant number of positive hits are left un-accessed through the traditional clone library screening system. In addition, these results also suggest that E. coli expression system might be intrinsically sub-optimal for screening for hemicellulases from environmental genomes compared to R. erythropolis system. The workflow required for screening one million clones in a fosmid library was estimated to be about 320 hours compared to 144 hours required via the mIVC-FACS screening platform. Some of the gene products obtained in both screening platforms show multiple substrate activities, suggesting that the microbial consortia of composting material consist of microorganisms that produce enzymes with multiple lignocellulytic activities. While this platform still requires optimisation, we have demonstrated that this technique can be used to isolate genes encoding enzymes from mixed microbial genomes. mIVC-FACS is a promising technology with the potential to take metagenomic studies to the second generation of novel natural products bio-prospecting. The astonishing sensitivity and ultra-high throughput capacity of this technology offer numerous advantages in metagenomic bio-prospecting. / National Research Foundation (NRF)

Les zoonoses constituent plus des deux tiers des pathologies virales qui concernent l’homme. Le développement et la démocratisation des outils de métagénomique en font de bons outils d’inventaire et de surveillance de virus potentiellement émergents.Dans un premier temps j’ai développé et validé un protocole expérimental de purification des viromes à ARN qui permettait le maintien de l’infectivité des particules virales. Ce protocole a ensuite été appliqué pour caractériser les communautés virales d’arthropodes hématophages et de prélèvements de faune sauvage. J’ai par la suite réalisé l’inventaire des communautés virales de viande de singe fumée illégalement importée en France et confisquée par les douanes, qui a révélé la présence de nombreux bactériophages, dont certains pourraient infecter des bactéries potentiellement pathogènes pour l’homme.Enfin j’ai caractérisé les communautés virales de culicoïdes collectés au Sénégal, ce qui a permis de mettre en évidence la présence de nombreux virus géants à ADN infectant les amibes. Le séquençage des viromes à ARN a quant à lui révélé la présence d'un certain nombre d'arbovirus qui pourraient constituer un risque d’émergence pour la santé humaine. Du fait de nombreux facteurs intrinsèques et extérieurs à l’agent infectieux, la prédiction des futures émergences de virus zoonotiques est très compliquée voire utopique, mais elle reste un challenge crucial et d’actualité. La stratégie de réalisation d’inventaires des communautés virales présentes dans les différents acteurs des cycles de transmission zoonotique est un premier pas indispensable dans la connaissance des risques potentiels d’émergence en population humaine. / Zoonoses are responsible of more than two thirds of human viral infections. The development of high-throughput sequencing tools and their application in metagenomics allow inventorying the viral communities of various reservoirs in order to detect the emergence of viruses before their infection to humans. In this context, I characterized the viral communities of simian bushmeat illegally imported into France and of Culicoides biting midges, recognized vectors of several viruses of human and veterinary medicine importance. I have first developed a protocol for the purification of RNA viromes which allowed maintaining the infectivity of viral particles. This protocol was subsequently applied to characterize viral communities of bloodsucking arthropods and wildlife samples. In a second part I realized the inventory of viral communities of smoked simian bushmeat illegally imported into France and confiscated by the French customs. This study revealed the presence of a wide diversity of bacteriophages, in which some of them could infect bacteria potentially pathogenic for humans.Finally I characterized the viral communities of Culicoides biting midges collected in Senegal, which revealed the presence of sequences related to several giant DNA viruses infecting amoeba. Sequencing of the RNA virome revealed the presence of several arboviruses that could constitute a risk of emergence of zoonoses for humans.The prediction of future emerging zoonotic viruses is very difficult, if not impossible. However the characterization of viral communities present in the different actors of zoonotic transmission cycle is a first step to evaluate potential risks of transmission to humans.

Les virus sont les entités biologiques les plus abondantes et diversifiées sur Terre et leur diversité est encore très peu connue. Récemment, la métagénomique virale a facilité l'exploration de cette diversité. Néanmoins, la plupart des viromes environnementaux générés à ce jour proviennent de régions tempérées et la plupart des viromes humains proviennent d’échantillons de selles, sang ou de prélèvements oro-naso-pharyngés. L'objectif de mon travail de thèse était d’apporter de nouvelles connaissances sur les communautés virales d’environnements et d’échantillons humains les moins étudiés en utilisant une approche de métagénomique virale.La première partie de cette thèse est une revue des principaux outils d'analyse des métagénomes viraux. La deuxième partie présente la première étude de métagénomique virale dans le désert du Sahara. Dans la troisième partie de ma thèse, je présente de viromes associés a l'Homme: i) le premier métagénome viral issu d'un coprolithe humain du Moyen Âge; ii) la première étude de métagénomique virale sur de liquides péricardiques provenant de patients atteints d’une péricardite infectieuse d'origine inconnue; iii) une analyse fonctionnelle de métagénomes viraux précédemment publiés associés aux expectorations de patients atteints de mucoviscidose qui décrit les gènes de résistance aux antibiotiques portés par les bactériophages dans ces patients.Ce travail présente ainsi des données inédites sur certaines communautés virales peu étudiées et confirme le potentiel de la métagénomique virale pour étudier la diversité virale, révéler la présence de virus inattendus ou inconnus et comprendre leur rôle dans leur écosystème d’origine. / Viruses are the most abundant and diverse organisms but little is known about their diversity. Recently, viral metagenomics has allowed performing broad unselective exploration of uncultivated viral communities, bypassing the limits of classical viral detection tools. However, most viral metagenomes are generated from temperate regions (for environmental studies) or from modern stool samples, sera/blood and naso-/oro- pharyngeal samples (for human-associated studies). Therefore, the purpose of my thesis is to study viral communities in the least investigated environments or human samples, using viral metagenomics.The first part of my thesis is a review of the main computational tools for the analysis of viral metagenomes. The second part of my thesis presents the first viromes generated from the Sahara desert. In the third part, I investigate human-associated viral communities: i) the first virome from a human coprolite; ii) the first viromes generated from human pericardial fluids, in idiopathic pericarditis cases; ii) a functional-level investigation of previously described viral metagenomes from cystic fibrosis patient sputa that focuses on antimicrobial resistance genes carried by bacteriophages to better understand the emergence of multidrug-resistance bacteria in the airways of cystic fibrosis patients.This thesis work provides original data on unexplored viral communities and shows the potential of viral metagenomics to give insights on viral diversity, reveal the presence of expected and unexpected viruses and decipher their role in the ecosystem.

Les glucides sont très rependus dans la nature et sont impliqués dans une multitude de phénomènes biologiques. Sous forme de saccharides et de glycoconjugués, ils constituent une partie substantielle de la biomasse produite sur terre et représentent une source potentielle d’énergie renouvelable de première importance. La diversité des glucides complexes est créée et contrôlée par un panel d’activités enzymatiques qui interviennent dans leur assemblage, dégradation et modification. L’étude structurale et fonctionnelle des enzymes actives sur les glucides (CAZymes) est à la base de multiples efforts de recherche appliquée en biotechnologie. L’industrie recherche actuellement des enzymes avec des activités et des spécificités encore plus performantes. L’activité de recherche de ces nouvelles enzymes est grandement facilitée par l’accumulation de séquences biologiques dans les bases de données, provenant notamment des études génomiques.Mon sujet de recherche s’inscrit dans un objectif de développement d’outils pour la classification et l’identification de nouvelles enzymes impliqués dans la conversion de la biomasse. Tous ces travaux sont en lien direct avec la mise en place d’une nouvelle infrastructure de la base de données CAZy et l’analyse de données génomiques, métagénomiques et biochimiques. La refonte complète de la structure de la base de données préexistantes et de son interface a été ainsi réalisée. Cet effort a été validé par l’analyse des familles de polysaccharide lyases et la création de sous-familles, dont l’homogénéité fonctionnelle a été révélée. De plus, la détection systématique de protéines modulaires portant des modules d’adhésion aux composants de la paroi végétale a permis l’identification de nouvelles protéines potentiellement impliquées dans la dégradation de la biomasse végétale. Enfin, j’ai implémenté des approches automatisées capables d’analyser de grands volumes de données (méta)génomiques pour en extraire le contenu en CAZymes. / Carbohydrates are widely distributed in nature, where they are involved in a multitude of important biological events. Saccharides and glycoconjugates constitute the main component of the biomass produced on earth, therefore they represent a plentiful source of renewable energy. The diversity of complex carbohydrates is created and controlled by a panel of enzyme activities involved in their assembly, degradation and modification. The structural and functional study of Carbohydrate Active enZymes on (CAZymes) has been the basis for many applied research efforts in biotechnology. For exemple, the biotechnology industry is currently searching enzymes with enhanced activities and specificities. The identification of new enzymes is potentially facilitated by the large-scale accumulation of gene sequences, particularly from current genomic studies.This thesis aimed at developing tools for the classification and identification of new enzymes involved in biomass degradation. To this end, a new structure of the CAZy database was developed and applied to mining genomic, metagenomic and biochemical data. A complete reorganisation of the structure of the existing database and its interface has been achieved. In this effort the analysis of all known families of polysaccharide lyases has been validated and subfamilies were created, which revealed functional homogeneity. In addition, the systematic identification of modular proteins containing plant cell wallbinding modules allowed the identification of new proteins potentially targeting plant biomass. Finally, I show that it is indeed possible to analyze large volumes of (meta)genomic data by automated methods in order to understand their CAZyme contents.

Orientadores: Valéria Maia Merzel, Itamar Soares de Melo / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Biologia / Made available in DSpace on 2018-08-26T12:19:42Z (GMT). No. of bitstreams: 1 Domingos_DanielaFerreira_D.pdf: 4459680 bytes, checksum: c21b44008997f453425f3732e5729029 (MD5) Previous issue date: 2014 / Resumo: O manguezal é um ambiente rico em diversidade microbiana, entretanto existem poucos estudos sobre esse tema no Brasil, tornando imperativo o conhecimento e a exploração de novos micro-organismos e seus metabólitos neste ecossistema. A prospecção da diversidade microbiana em ambientes pouco explorados, como os manguezais, potencializa as chances de sucesso na busca por novas moléculas bioativas, contribuindo para o desenvolvimento econômico e ambiental mais sustentável. Entretanto, sabemos que embora as técnicas de cultivo tenham sido aprimoradas e tenham permitido a recuperação in vitro de um número crescente de micro-organismos ainda não cultivados, nosso conhecimento sobre sua ecologia permanece insuficiente para cultivar a maioria deles. Neste contexto, as bibliotecas metagenômicas surgem como uma ferramenta poderosa para acessar de maneira mais abrangente a diversidade microbiana total em um dado ambiente, permitindo a análise e exploração de genes funcionais de membros da microbiota, principalmente de micro-organismos não-cultivados, e a descoberta de novos compostos bioativos. O objetivo geral desse trabalho foi a prospecção de biossurfactantes a partir de microbiota de manguezal utilizando uma abordagem polifásica. Para a abordagem independente-de-cultivo, foi construída uma biblioteca metagenômica de alto peso molecular a partir de sedimento de manguezal contaminado com petróleo. Os clones obtidos foram submetidos à triagem funcional e molecular para compostos com atividade biossurfactante. Três clones potencialmente produtores foram selecionados e submetidos ao sequenciamento fosmidial para a caracterização gênica dos insertos. Embora os clones apresentassem uma redução na tensão superficial, não foram identificados genes que pudessem estar envolvidos na síntese de algum biossurfactante. Os resultados obtidos revelaram que o uso da metagenômica funcional para a exploração metabólica de uma comunidade microbiana pode oferecer grandes limitações quando se trata de prospecção de biossurfactantes, cujos operons são muitas vezes maiores que 30-40 kb e podem conter elementos de regulação esparsos no genoma da bactéria selvagem. Na abordagem dependente-de-cultivo, foram selecionadas duas linhagens produtoras de biossurfactantes, Bacillus safensis CCMA-560 e Gordonia sp. CCMA-559, isoladas e testadas em estudo prévio. Através de planejamento experimental do tipo Plackett-Burman e Delineamento Composto Central Rotacional foi otimizada a produção do biossurfactante por essas linhagens. A pumolicidina produzida pelo B. safensis CCMA-560 foi caracterizada quimicamente, bem como a via metabólica responsável por sua produção. Este foi o primeiro trabalho a reportar a análise fisiológica, genética e química da produção de biossurfactante por representantes da espécie B. safensis / Abstract: Mangrove is an environment rich in microbial diversity, but little has been reported about it in Brazil, making the knowledge and exploitation of new microorganisms in mangroves an imperative issue. The prospection of microbial diversity in underexplored environments, such as mangroves, increses possibility of success in looking for new bioactive molecules, contributing to a more sustentable economy and environment. Although the cultivation techniques have been improved, allowing an increased number of yet uncultivated microorganisms to be able recuperated in vitro, our knowledge about their ecology is still insufficient to cultivate the majority of them. Therefore, the metagenomic library have emerged as a powerful tool to more widely access the overall microbial diversity in an environment, enabling the functional analyses of genes and the discovery of new bioactive compounds, mainly uncultured-microorganisms. The aim of the current work was the prospection of biosurfactant from mangrove microbiota through a polifasica approach. For the independent-cultivation approach, a metagenomic library of high molecular weight was constructed from mangrove sediment contaminated with oil. The clones were subjected to functional and molecular screening to identify biosurfactant activity. Three clones with the potential to procuce biosurfactant were selected, and the fosmid sequencing for genetic chatacterization of the insert was carried out. Although the clones showed the ability to reduce the surface tension, genes that may involved in biosurfactant synthesis were not identified. The results showed that using the functional metagenomic to explore metabolic pathways of microbial communities can have great limitation when in the prospection of biosurctants when the operon are larger then 30-40 kb and the genome of wild bacteria contains sparse regulation elements. For the dependent-cultivation approach, two strains that produce biosurfactant were selected: Bacillus safensis CCMA-560 and Gordonia sp. CCMA-559. They were isolated and tested in previous study. The biosurfactant production was opmized through the Placktt-Burman design and the Central Composite Design. The pumilacidin produced by B. safensis CCMA-560 was chemically characterized, and the metabolic pathway that is responsible for its production was genetically characterized. This work was the first report the physiologic, genetic, and chemical analysis on the biosurfactant production by the B. safensis strain / Doutorado / Genetica de Microorganismos / Doutora em Genética e Biologia Molecular

One in 3,000 people in the US are born with cystic fibrosis (CF), a genetic disorder affecting the reproductive system, pancreas, and lungs. Lung disease caused by chronic bacterial and fungal infections is the leading cause of morbidity and mortality in CF. Identities of the microbes are traditionally determined by culturing followed by phenotypic and biochemical assays. It was first thought that the bacterial infections were caused by a select handful of bacteria such as S. aureus, H. influenzae, B. cenocepacia, and P. aeruginosa. With the advent of PCR and molecular techniques, the polymicrobial nature of the CF lung became evident. The CF lung contains numerous bacteria and the communities are diverse and unique to each patient. The total complexity of the bacterial infections is still being determined. In addition, only a few members of the fungal communities have been identified. Much of the fungal community composition is still a mystery. This dissertation addresses this gap in knowledge. A snap shot of CF sputa bacterial community was obtained using the length heterogeneity-PCR community profiling technique. The profiles show that south Florida CF patients have a unique, diverse, and dynamic bacterial community which changes over time. The identities of the bacteria and fungi present were determined using the state-of-the-art 454 sequencing. Sequencing results show that the CF lung microbiome contains commonly cultured pathogenic bacteria, organisms considered a part of the healthy core biome, and novel organisms. Understanding the dynamic changes of these identified microbes will ultimately lead to better therapeutical interventions. Early detection is key in reducing the lung damage caused by chronic infections. Thus, there is a need for accurate and sensitive diagnostic tests. This issue was addressed by designing a bacterial diagnostic tool targeted towards CF pathogens using SPR. By identifying the organisms associated with the CF lung and understanding their community interactions, patients can receive better treatment and live longer.

Cubatão, o maior pólo industrial da américa latina também já foi uma das cidades mais poluídas do mundo. Os 30 anos de intensa atividade industrial vêm pressionando o meio ambiente com substâncias tóxicas e afetando gravemente a saúde da população. Dentre as substâncias contaminantes mais importantes da região estão os derivados de petróleo como o benzeno, tolueno, etilbenzeno e xilenos. Conhecidos como BTEX, eles são produzidos e utilizados em larga escala e a contaminação ocorre frequentemente através de vazamentos. Nos solos, devido à sua solubilidade em água, essas substâncias podem se espalhar por longas distâncias a partir do ponto afetado contaminando locais distantes. Já foi comprovada a capacidade de micro-organismos de sobreviver e até utilizar BTEX como fonte de carbono. Os micro-organismos adaptados catabolizam os contaminantes transformando-os em substâncias menos tóxicas e até mesmo eliminando-os do ambiente, capacidade de grande interesse econômico e ambiental. Nessa linha, nossa proposta visa o estudo das comunidades microbianas de solos afetados e não afetados por BTEX. Para isso foi utilizada a metagenômica como abordagem de estudo identificando-se diferenças qualitativas e quantitativas nas estruturas microbianas de três diferentes locais do município de Cubatão, sendo um deles afetado diretamente por BTEX. Pelo método utilizado e aqui desenvolvido, foi possível identificar um panorama metabólico geral identificando-se genes relevantes e o potencial de degradação de hidrocarbonetos aromáticos de micro-organismos conhecidos e desconhecidos, revelando melhor o potencial metabólico dos solos identificados. Os resultados apresentados podem contribuir para um melhor entendimento da dinâmica in situ de uma comunidade microbiana afetada por BTEX assim como melhorar o conhecimento sobre a comunidade microbiana de um local altamente impactado como Cubatão. / Cubatão is the largest industrial site in Latin America and was in the past one of the most polluted cities in the world. 30 years of intense industrial activity has pushed environmental limits with toxic substances and has severely affected the inhabitants\' health. Among the contaminants found in the region, the petroleum derivatives benzene, toluene, ethylbenzene and xylenes are the most important. Known collectively as BTEX, they are produced and used at a large scale and contamination frequently occurs. Because it is highly soluble in water, when in soil BTEX can spread long distances from the original contamination site, thus affecting large areas. Some microorganisms are known to live in contaminated environments and use contaminants such as BTEX as a unique carbon source for energy production. They catabolize contaminants into less dangerous products or even eliminate them from environment, a feature which has great commercial and environmental interest. We therefore compared the microbial communities in soils which were affected and un-affected by BTEX contamination. To this end, we used a metagenomics approach and developed a comparison method to identify microorganisms and degradation potential of soils studied. We found qualitative and quantitative differences in microbial structures from three different sites in Cubatão County, one of which is contaminated with BTEX. We constructed a metabolic overview identifying important genes, degradation potential and microorganisms related to BTEX degradation. The results presented here could contribute to understanding the in situ dynamics of a BTEX affected microbial community as well as improving our knowledge of the microbial community of Cubatão, a highly environmentally impacted place.

Este trabalho teve como objetivo caracterizar a diversidade bacteriana presente em uma área de mineração. Os resultados mostraram que o filo mais abundante das bactérias classificadas foi o proteobacteria e 66% deste filo pertencem a classe betaproteobacteria. A diversidade de bactérias encontradas no solo de fora da mina mostrou-se diferente daquele encontrado dentro da mina, mostrando que a ação da mineração poderia ter modificado a comunidade microbiana nativa do local. Na tentativa de elucidar o mecanismo de resistência ao cobre presente nos isolados, foram escolhidas 2 cepas resistentes a altas concentrações de cobre: Acinetobacter sp. e Enterobacter sp. Os resultados de proteômica mostraram que as duas cepas possuem mecanismos diferentes de resistência. Acinetobacter teve a expressão de duas proteínas CopA e CopB em diferentes condições de cultivos (200 ppm de cobre 6h de cultivo; 24h e 72h de cultivo em 450 ppm de cobre). Com esses resultados foi possível elucidar que esta cepa tem o mesmo mecanismo de resistência encontrado em Enterococcus hirae, onde a proteína CopA tem como função o influxo de cobre e a proteína CopB o efluxo do excesso de cobre intracelular. Já a bactéria Enterobacter teve expressão apenas de proteínas de membranas e nenhuma proteína propriamente conhecida como de resistência à cobre. Este resultado foi corroborado por técnica de PCR, no qual não houve expressão para esse gene. Possivelmente as proteínas de membrana externa estão responsáveis pela homeostase do cobre realizada por esta cepa. / This work aimed to characterize the bacterial diversity present in a mining area. Results showed the most abundant phylum of the classified bacteria are proteobacteria and 66% of this phylum are class betaproteobacteria. The diversity of bacteria found in the soil outside the mine differed from those found within the mine, showing that the mining action may have modified the local microbial community. In an attempt to elucidate the mechanism of resistance to copper present in the isolates, two strains resistant to high concentrations of copper were chosen: Acinetobacter sp. and Enterobacter sp. Proteomics results showed these two strains have different mechanisms of resistance. Acinetobacter had the expression of two CopA and CopB proteins under different cultivation conditions (200 ppm of copper 6h of culture and 24h and also in 450 ppm of copper and 72h of culture). With these results it was possible to elucidate, this strain has the same mechanism of resistance found in Enterococcus hirae, which the CopA protein functions as the copper influx and the CopB protein the efflux of excess intracellular copper. On the other hand, Enterobacter bacterium had only membrane proteins, and no proteins known as copper resistance. This result was corroborated with PCR, where there was no expression for this gene. Possibly the outer membrane proteins are responsible for the copper homeostasis performed by this strain.

Vinyl chloride (VC) is a human carcinogen and common groundwater contaminant in the United States. Some of the indigenous bacteria can utilize VC for their growth, which is important for bioremediation. As previous studies have been majorly focused on VC-degrading bacteria in pure cultures, we initiated the study to investigate the microbial community structure and interactions in more complicated systems, such as mixed-pure cultures and groundwater enrichment cultures. Finally, we extended our study into the field by investigating the diversity and abundance of functional genes in VC-assimilating pathways at six contaminated sites. In our first study, Nocardioides was found to be the most dominant genus in Carver groundwater enrichment cultures via stable-isotope probing and 16S rRNA gene amplicon Illumina sequencing. As cross-feeding was observed, in the second study, mixed-pure culture experiment was conducted to explore the potential effects of VC-assimilating Nocardioides on other bacteria, which showed VC cometabolizer Mycobacterium strain JS622 would take up carbon from VC to sustain their growth when mixed with VC-assimilating Nocardioides sp. strain JS614. The third study was conducted with a different groundwater source from Fairbanks, AK, which again showed Nocardioides is dominant in the microbial community. A novel VC-assimilating Nocardioides sp. bacteria was isolated, named XL1. The putative genome of XL1 extracted from enrichment culture metagenome was 99% to 100% identical to strain JS614, with a plasmid genome bin similar to strain JS614 plasmid pNOCA01, though the morphology of strain XL1 was distinct from strain JS614. About 90% of the plasmid contigs could be mapped onto Nocardioides sp. strain JS614 plasmid with 100% identity, containing known aerobic ethene and VC degradation pathway genes encoding alkene monooxygenase and epoxyalkane: coenzyme M transferase (EaCoMT). Glutathione synthase and glutathione S-transferase genes, possibly involved in epoxide detoxification, were found in Polaromonas, Mesorhizobium and Pseudomonas-affiliated genome bins. The study also showed cultures adapted to VC faster after amended with ethene. The in-situ study (the fourth study) revealed 192 different EaCoMT T-RFs from six chlorinated ethene contamination sites via T-RFLP analysis, implicating higher EaCoMT diversity than previously known. Phylogenetic analysis revealed that a majority of the 139 cloned sequences (78.4%) grouped with EaCoMT genes found in VC- and ethene-assimilating Mycobacterium strains and Nocardioides sp. strain JS614. EaCoMT gene abundance was positively correlated with VC and ethene concentrations at the sites studied.

Les progrès technologiques des vingt dernières années ont permis l’avènement d'une biologie à haut-débit reposant sur l'obtention de données à grande échelle de façon automatique. Les statisticiens ont un rôle important à jouer dans la modélisation et l'analyse de ces données nombreuses, bruitées, parfois hétérogènes et recueillies à différentes échelles. Ce rôle peut être de plusieurs natures. Le statisticien peut proposer de nouveaux concepts ou méthodes inspirées par les questions posées par cette biologie. Il peut proposer une modélisation fine des phénomènes observés à l'aide de ces technologies. Et lorsque des méthodes existent et nécessitent seulement une adaptation, le rôle du statisticien peut être celui d'un expert, qui connaît les méthodes, leurs limites et avantages. Le travail présenté dans cette thèse se situe à l'interface entre mathématiques appliquées et biologie, et relève plutôt des deuxième et troisième type de rôles mentionnés.Dans une première partie, j’introduis différentes méthodes développées pour l'analyse de données biologiques à haut débit, basées sur des modèles à variables latentes. Ces modèles permettent d'expliquer un phénomène observé à l'aide de variables cachées. Le modèle à variables latentes le plus simple est le modèle de mélange. Les deux premières méthodes présentées en sont des exemples: la première dans un contexte de tests multiples et la deuxième dans le cadre de la définition d'un seuil d'hybridation pour des données issues de puces à ADN. Je présente également un modèle de chaînes de Markov cachées couplées pour la détection de variations du nombre de copies en génomique prenant en compte de la dépendance entre les individus, due par exemple à une proximité génétique. Pour ce modèle, nous proposons une inférence approchée fondée sur une approximation variationnelle, l'inférence exacte ne pouvant pas être envisagée dès lors que le nombre d'individus augmente. Nous définissons également un modèle à blocs latents modélisant une structure sous-jacente par bloc de lignes et colonnes adaptées à des données de comptage issue de l'écologie microbienne. Les données issues de méta-codebarres ou de métagénomique correspondent à l'abondance de chaque unité d'intérêt (par exemple micro-organisme) d'une communauté microbienne au sein d'environnement (rhizosphère de plante, tube digestif humain, océan par exemple). Ces données ont la particularité de présenter une dispersion plus forte qu'attendue sous les modèles les plus classiques (on parle de sur-dispersion). La classification croisée est une façon d'étudier les interactions entre la structure des communautés microbiennes et les échantillons biologiques dont elles sont issues. Nous avons proposé de modéliser ce phénomène à l'aide d'une distribution Poisson-Gamma et développé une autre approximation variationnelle pour ce modèle particulier ainsi qu'un critère de sélection de modèle. La flexibilité et la performance du modèle sont illustrées sur trois jeux de données réelles.Une deuxième partie est consacrée à des travaux dédiés à l'analyse de données de transcriptomique issues des technologies de puce à ADN et de séquençage de l’ARN. La première section concerne la normalisation des données (détection et correction de biais techniques) et présente deux nouvelles méthodes que j’ai proposées avec mes co-auteurs et une comparaison de méthodes à laquelle j’ai contribuée. La deuxième section dédiée à la planification expérimentale présente une méthode pour analyser les dispositifs dit en dye-switch.Dans une dernière partie, je montre à travers deux exemples de collaboration, issues respectivement d'une analyse de gènes différentiellement exprimés à partir de données issues de puces à ADN, et d'une analyse du traductome chez l'oursin à partir de données de séquençage de l'ARN, la façon dont les compétences statistiques sont mobilisées et la plus-value apportée par les statistiques aux projets de génomique. / The technological progress of the last twenty years allowed the emergence of an high-throuput biology basing on large-scale data obtained in a automatic way. The statisticians have an important role to be played in the modelling and the analysis of these numerous, noisy, sometimes heterogeneous and collected at various scales. This role can be from several nature. The statistician can propose new concepts, or new methods inspired by questions asked by this biology. He can propose a fine modelling of the phenomena observed by means of these technologies. And when methods exist and require only an adaptation, the role of the statistician can be the one of an expert, who knows the methods, their limits and the advantages.In a first part, I introduce different methods developed with my co-authors for the analysis of high-throughput biological data, based on latent variables models. These models make it possible to explain a observed phenomenon using hidden or latent variables. The simplest latent variable model is the mixture model. The first two presented methods constitutes two examples: the first in a context of multiple tests and the second in the framework of the definition of a hybridization threshold for data derived from microarrays. I also present a model of coupled hidden Markov chains for the detection of variations in the number of copies in genomics taking into account the dependence between individuals, due for example to a genetic proximity. For this model we propose an approximate inference based on a variational approximation, the exact inference not being able to be considered as the number of individuals increases. We also define a latent-block model modeling an underlying structure per block of rows and columns adapted to count data from microbial ecology. Metabarcoding and metagenomic data correspond to the abundance of each microorganism in a microbial community within the environment (plant rhizosphere, human digestive tract, ocean, for example). These data have the particularity of presenting a dispersion stronger than expected under the most conventional models (we speak of over-dispersion). Biclustering is a way to study the interactions between the structure of microbial communities and the biological samples from which they are derived. We proposed to model this phenomenon using a Poisson-Gamma distribution and developed another variational approximation for this particular latent block model as well as a model selection criterion. The model's flexibility and performance are illustrated on three real datasets.A second part is devoted to work dedicated to the analysis of transcriptomic data derived from DNA microarrays and RNA sequencing. The first section is devoted to the normalization of data (detection and correction of technical biases) and presents two new methods that I proposed with my co-authors and a comparison of methods to which I contributed. The second section devoted to experimental design presents a method for analyzing so-called dye-switch design.In the last part, I present two examples of collaboration, derived respectively from an analysis of genes differentially expressed from microrrays data, and an analysis of translatome in sea urchins from RNA-sequencing data, how statistical skills are mobilized, and the added value that statistics bring to genomics projects.