Single-cell RNA sequencing (scRNA-seq) has been a powerful technique for characterizing transcriptional heterogeneity related to tumor development and disease pathogenesis. Despite the advances of technology, there is still a lack of software to systematically and easily assess the quality and different types of artifacts present in scRNA-seq data and a statistical framework for understanding heterogeneity in the gene programs of cancer cells. In this dissertation, I first introduced novel computational software to enhance and streamline the process of quality control for scRNA-seq data called SCTK-QC. SCTK-QC is a pipeline that performs comprehensive quality control (QC) of scRNA-seq data and runs a multitude of tools to assess various types of noise present in scRNA-seq data as well as quantification of general QC metrics. These metrics are displayed in a user-friendly HTML report and the pipeline has been implemented in two cloud-based platforms. Most scRNA-seq studies only profiled a small number of tumors and provided a narrow view of the transcriptome in tumor tissue. Next, I developed a novel framework to perform a large-scale meta-analysis of cancer cells from 12 studies with scRNA-seq data from patients with non-small-cell lung cancer (NSCLC). I discovered interpretable gene co-expression modules with celda and demonstrated that the activity of gene modules accounted for both inter- and intra-tumor heterogeneity of NSCLC samples. Furthermore, I used CaDRa to determine that the levels of some gene modules were significantly associated with combinations of underlying genetic alterations. I also showed that other gene modules are associated with immune cell signatures and may be important for communication with the cancer cells and the immune microenvironment. Finally, I presented a novel computational method to study the association between copy number variation (CNV) and gene expression at the single-cell level. The diversity of the CNV profile was identified in tumor subclones within each sample and I discovered cis and trans gene signatures which have expression values associated with specific somatic CNV status. This study helped us prioritize the potential cancer driver genes within each CNV region. Collectively, this work addressed the limitation in the quality control of scRNA-seq data and provided insights for understanding the heterogeneity of NSCLC samples.

The variability in calf management can change the physiological state of the calf as they are weaned or attain puberty. It is up to the producer to ensure that the calves develop properly to meet their expected needs on the farm. While there are guidelines from the NRC in place, there is a substantial range in the amount of protein and fat that a calf can be fed. This physiological state can be reflected in the proteins produced in tissues, the expression of gene regulatory pathways, or even the microbes present in the gut. The purpose of this study was to examine how an increase in dietary energy in milk replacer of pre-weaned Holstein heifers impacts the microbial profile of the rumen as well as the transcriptome in tissues related to growth and metabolism. Our hypothesis was that pre-weaned Holstein heifers on milk replacer diets with lower dietary energy will have a different rumen microbiome composition and a different transcriptome in growth related tissues. Holstein heifer calves (n = 36) were assigned randomly to 1 of 2 milk replacer diets: restricted (R; 20.9% CP, 19.8% Fat; n = 18) or enhanced (E; 28.9% CP, 26.2% Fat; n = 18). Calves were euthanized and rumen fluid was collected at pre-weaning (8 wks; n = 6) or post-weaning (10 wks; n = 6). Liver (L), adipose (A), and longissimus dorsi (LD) tissues were collected at pre-weaning (8 wks; n = 12). Average daily gain (ADG) and gain-to-feed ratio (G:F) were calculated for each calf. Analysis of ADG and G:F was performed using a PROC GLM in SAS with diet as the main effect; E calves had increased ADG and G:F compared to R calves. For rumen samples, libraries were constructed from extracted DNA and DNASeq was conducted using a paired-end analysis at 100 bp using Illumina HiSeq 2500. Operational taxonomic unit (OTU) clustering analysis was conducted using the 16s rRNA Greengenes reference. A PERMANOVA analysis was conducted in R to determine OTU populations for age and treatment. There was no difference in microbiome composition between pre-weaning and post-weaning calves (P = 0.761). Microbiome composition differed between E and R calves (P < 0.001). Bacteroidetes and Firmicutes represented the most abundant phyla for both E and R calves. Enhanced calves had 49.4% (5141 reads) Bacteriodetes and 36.4% (3789 reads) Firmicutes; whereas, R calves had 31.6% (2491 reads) Bacteriodetes and 41.1% (3236 reads) Firmicutes. For L, A, and LD samples, libraries were constructed from extracted RNA for RNA-Seq analyses. RNA-Seq analysis was performed using CLC Genomics Workbench and the Robinson and Smith Exact Test was used to identify differentially expressed genes between diets. There were 238 differentially expressed genes in A, 227 in LD, and 40 in L. Of the differentially expressed genes, 10 appeared in at least 2 tissues. PANTHER was used to identify functional categories of differentially expressed genes. The majority of genes were associated with metabolic processes (A = 112, 26.7%; L = 16, 32.0%; LD = 81, 34.0%) or cellular processes (A = 93, 22.1%; L = 13, 26.0%; LD = 73, 30.7%). In E calves, upregulated genes included those regulating NADH dehydrogenation (LD = 17, A = 5; i.e. ND1, ND4), gluconeogenesis (LD = 2, A = 6; i.e. ALDOB, PCK2), and cell proliferation (LD = 2, A = 3; i.e. GADD45A, CDKN1A). There was a difference in both the transcriptome and rumen microbiome of calves fed differing levels of dietary energy. The calves on the R diet had a rumen microbial composition more similar to a younger calf, while the composition of E calves was more similar to a mature calf. The change in regulation of genes involved in the cell cycle and ATP synthesis in response to dietary energy could explain the change in ADG between diets. Because the R calves appeared to have stunted development of their microbiomes and an expression profile similar to oxidative stress, it is possible that the R diet did not meet the nutritional requirements of that calves. / Master of Science / Changes in the way a calf is raised from birth can affect the biological processes that occur when they change from liquid to solid feed or reach reproductive maturity. While there are guidelines in place in how much a calf should be fed, there is still a large range in the amount of protein and fat in the liquid feed. The change in nutrition levels changes the biological processes occurring in the calf, which are reflect by changes in expression of genes in different parts of the calf as well the levels of microbes in the gut. The purpose of this study was to examine how the change in protein and fat in the liquid feed of female calves affects the microbes in the first section of the stomach, the rumen, as well as the genes expressed in parts of the calf associated with growth. Our hypothesis was that female calves fed liquid diets with lower protein and fat will have different rumen microbes and a different level of gene expression in growth related tissues. Female calves (n = 36) were randomly assigned 1 of 2 diets at birth: restricted (R; 20.9% Crude Protein, 19.8% Fat; n = 18) or enhanced (E; 28.9% Crude Protein, 26.2% Fat; n = 18). Calves were euthanized and rumen contents were collected at removal of the liquid feed (8 wks; n = 6) or 2 wks after calves were switched to an all dry feed diet (10 wks; n = 6). Liver (L), adipose (A), and longissimus dorsi (LD) tissues were collected at removal of the liquid feed (8 wks; n = 12). Bacterial DNA was extracted from the rumen samples and RNA was extracted from L, A, and LD samples. DNA and RNA were sequenced at the University of Missouri DNA Core Lab. Microbiome composition differed between E and R calves (P < 0.001). Enhanced calves had 49.4% Bacteriodetes and 36.4% Firmicutes; whereas, R calves had 31.6% Bacteriodetes and 41.1% Firmicutes. There were 238 differentially expressed genes in A, 227 in LD, and 40 in L. Of the differentially expressed genes, 10 appeared in at least 2 tissues. In E calves, upregulated genes included those regulating NADH dehydrogenation (LD = 17, A = 5; i.e. ND1, ND4), gluconeogenesis (LD = 2, A = 6; i.e. ALDOB, PCK2), and cell growth (LD = 2, A = 3; i.e. GADD45A, CDKN1A). There was a difference in both the gene expression and rumen microbiome of calves fed differing levels of protein and fat. The calves on the R diet had a rumen microbial composition more similar to a younger calf, while the composition of E calves was more similar to a mature calf. Because the R calves appeared to have stunted development of their microbiomes and an expression profile similar to oxidative stress, it is possible that the R diet did not meet the nutritional requirements of that calves.

Bacterial transcriptomes are highly complex, comprising not only protein-coding RNAs and translation-related non-coding RNAs, but also non-coding RNAs that function as regulators of gene expression. The post-transcriptional modification of RNA sequences by RNA editing enzymes, which has recently been shown to affect diverse RNA substrates in several bacteria, can magnify this complexity further still. However, little is known about RNA editing and non-coding RNAs in Streptomyces venezuelae, a model organism for studying complex bacterial development and specialized metabolism. This thesis investigates RNA editing and non-coding regulatory RNAs in S. venezuelae using RNA sequencing data from wild type and mutant strains at various stages of development and under several laboratory-controlled conditions. We identified hundreds of adenosine-to-inosine editing events throughout the transcriptome and predicted the potential impact of the edits occurring in protein-coding RNAs. The potential role of the adenosine deaminase enzyme TadA in facilitating these RNA editing events is also considered. Additionally, we detected thousands of transcripts that are expressed from unannotated regions of the S. venezuelae genome, many of which we predict are non-coding RNAs. Furthermore, we highlight our efforts to characterize a highly expressed putative non-coding RNA that exhibits considerable sequence conservation in other streptomycetes. This work provides new insights into the transcriptomic complexity of S. venezuelae and expands our understanding of RNA-based regulation in bacteria. / Thesis / Master of Science (MSc) / All living things have DNA, which contains the instructions for maintaining life in the form of genes. These genes are copied into RNAs, and some of these RNA molecules are used to make proteins, which are the building blocks and machinery of cells. However, not all RNAs make proteins; some act as regulators, controlling which genes and proteins are active. Additionally, some proteins edit the instructions contained by RNA molecules after they are made, adding another layer of complexity to how cells regulate their activities. This thesis investigates these processes in Streptomyces venezuelae, a soil-dwelling bacterium known for its complex development and metabolism. We found hundreds of cases where RNA molecules are edited, potentially affecting their functions in the cell, and discovered thousands of non-protein-coding RNAs that may regulate genes or proteins. Our findings expand our understanding of how Streptomyces bacteria manage their complex genetic activities at the RNA level.

Yes / Functional genomics research can give us valuable insights into bacterial gene function. RNA Sequencing (RNA-seq) can generate information on transcript abundance in bacteria following abiotic stress treatments. In this study, we used the RNA-seq technique to study the transcriptomes of the opportunistic nosocomial pathogen Pseudomonas aeruginosa PAO1 following heat shock. Samples were grown at both the human body temperature (37 C) and an arbitrarily-selected temperature of 46 C. In this work using RNA-seq, we identified 133 genes that are differentially expressed at 46 C compared to the human body temperature. Our work identifies some key P. aeruginosa PAO1 genes whose products have importance in both environmental adaptation as well as in vivo infection in febrile hosts. More importantly, our transcriptomic results show that many genes are only expressed when subjected to heat shock. Because the RNA-seq can generate high throughput gene expression profiles, our work reveals many unanticipated genes with further work to be done exploring such genes products. / University of Malaya High Impact Research (HIR) UM-MOHE HIR Grants (UM.C/625/1/HIR/MOHE/CHAN/14/1, No. H-50001-A000027; UM.C/625/1/HIR/MOHE/CHAN/01, No. A000001-50001); PPP Grant (PG081-2015B)

Ces dernières années de nombreuses recherches ont montré l’importance des petits ARN dans la régulation de l’expression des gènes, chez tous les organismes vivants, des bactéries aux mammifères. Le projet de cette thèse était de recherche et d’identifier des petits ARN chez une bactérie pathogène pour l’homme, Helicobacter pylori (Hp). Cette bactérie colonise exclusivement l’estomac humain, un organe qui pendant longtemps a été considéré comme étant stérile, en raison du pH parfois très acide qui y règne. L’infection persistante de l’estomac humain causée par Hp est associée avec plusieurs pathologies gastriques tels que les gastrites, les ulcères peptiques, les cancers gastriques et les lymphomes du MALT. La moitié de la population est infectée par Hp, qui est responsable d’environ 1 million de décès par an à travers le monde, et de 6000 nouveaux cas de cancers gastrique par an en France. Au cours de ma thèse, j’ai travaillé en étroite collaboration avec le groupe du Pr. Jörg Vogel (RNA Biology, MPI, Berlin, Allemagne) pour développer une méthode rapide et efficace d’analyse du transcriptome complet d’Hp, en s’appuyant sur une nouvelle sur une technologie émergente de pyroséquençage haut-débit (HTPS 454 technology, Life Science, USA). Notre méthode de séquençage du transcriptome d’Hp à partir de banques enrichies en transcrits primaires (dRNA-seq), nous a permis d’identifier les sites d’initiation de la transcription (TSS) de milliers de d’ARN. Plus de la moitié de ces TSS ont été associés à des petits ARN non codants, de courte taille (de 50 à 250 nucléotides en moyenne), qui n’avaient jamais été découverts jusqu’alors, et dont les gènes sont localisés dans des régions intergéniques (sRNA) ou en antisens (asRNA) par rapport aux ORF précédemment annotées dans le génome d’Hp. Nos travaux ont également permis de mettre en évidence une forte activité de transcription antisens sur l’ensemble du génome de la bactérie, un phénomène déjà observé chez E. coli et les eucaryotes. Ainsi, au moins un TSS est localisé sur le brin opposé à 46 % des ORF et à 28% des régions « leaders » des précurseurs des ARNr 23S et 16S, et des ARNt. Enfin, l’approche dRNA-seq a permis l’identification de la première famille de toxines de type I (AapA) identifiée à ce jour chez Hp. Dans ces conditions normales de culture, la traduction de ces toxines est constitutivement réprimée par des petits ARN antisens (IsoA) qui ciblent les ARNm aapA par complémentarité de base. Malgré leur homologie avec des modules toxine-antitoxine identifiés chez d’autres bactéries, pour certaines impliquées dans la réponse aux stress, nous n’avons pas encore découvert les conditions dans lesquelles ces peptides aapA seraient exprimées chez Hp, et leur rôle biologique reste à élucider. / In the past few years, small regulatory RNAs have emerged as an important class of post-transcriptional regulators of gene expression. Indeed they have been identified and/or predicted to exist in all species ranging from bacteria to mammals. The project of this thesis was to search for small non coding RNAs in a human pathogen: Helicobacter pylori (Hp). This bacterium exclusively colonizes the human stomach, an organ that until recently was thought to be sterile due to its extreme acidity. It is now established that persistent colonization by Hp is associated with various gastric pathologies including gastritis, peptic ulcer, gastric cancer and MALT lymphoma. Half of the human population is infected by Hp that is responsible for about 1 million deaths per year and around 6000 cases of gastric cancer in France. During my thesis we , in a close collaboration with the group of Joerg Vogel (RNA biology, MPI, Berlin, Germany) developed a rapid and efficient method to reveal the whole transcriptome of Hp based on recent advances in high-throughput pyrosequencing technologies (HTPS 454 technology, Life Science, USA). By using specifically enriched libraries in primary transcripts, our strategy allowed us to map thousand (1907) of transcription start sites (TSS) on the Hp genome. More than half of these TSS correspond to new short transcripts (non coding RNAs, between 50 and 250 nucleotides in length) that have never been annotated in this genome and that are localized both in intergenic regions (sRNA) and in regions antisense to annotated ORFs (asRNA). Analysis of associations between primary transcription start sites (pTSS) revealed more complexity in the Hp transcriptome than previously anticipated: around one third (27%) of pTSS belong to antisense transcripts (aTSS). The strikingly high degree of antisense transcription occurs, similar to E. coli and higher eukaryotes, across the entire Hp genome. Overall, at least one aTSS is linked to ~46% of all ORFs, ~28% of tRNAs, and the 5’ leaders of 23S and 16S rRNA precursors. Finally our dRNA-seq approach led us to identify the first family of putative type I toxins (AapA) in the Hp genome. Under normal growth conditions these toxins are constitutively repressed by a sophisticated antisense RNA-mediated (IsoA) mechanism. Despite their homology to other toxin-antitoxin modules previously described in other bacteria, we have not found physiological conditions under which these peptides are expressed and have yet to determine the biological significance (if any ?) of these suicide genes.

La maladie de la rouille foliaire du peuplier, causée par le basidiomycète Melampsora larici-populina (Mlp) cause des dégâts importants dans les peupleraies européennes. Le séquençage du génome de la souche 98AG31 de Mlp a ouvert de nouvelles perspectives pour l'identification de déterminants géniques impliqués dans le processus infectieux du champignon et notamment ceux codant des effecteurs fongiques capables de manipuler la structure et le fonctionnement de la cellule hôte pour assurer le succès de l'infection. L'analyse du transcriptome du champignon au cours des différentes phases du processus infectieux, basée sur l'utilisation de puces à oligonucléotides NimbleGen ou le séquençage massifs d'ESTs, a permis d?identifier des gènes marqueurs de la germination, de la phase de croissance biotrophe et de la sporulation du champignon. Nous avons notamment pu montrer l'induction importante de nombreux gènes codant des petites protéines sécrétées (SSPs) au cours de la phase biotrophe à 96 hpi heures post-inoculation (hpi) ainsi qu'au sein du parenchyme lacuneux à 168 hpi par microdissection à capture laser. L?analyse fine du sécrétome de Mlp, basée sur l'annotation, l'évolution et l'expression des gènes codant des SSPs a permis de mettre à jour des effecteurs candidats. Certains, spécifiquement exprimés in planta ou présentant des homologies de séquence avec des effecteurs de rouilles ont été localisés au niveau de l'haustorium. De manière intéressante, d'autre gènes candidats appartenant à des familles multigéniques sous pression de sélection positive, sont riches en cystéines, spécifiquement exprimés in planta et possèdent un motif de translocation potentiellement impliqué dans l'export de l'effecteur dans la cellule hôte. Ce travail d'analyse fine des effecteurs potentiels d'un agent de rouille à l'échelle génomique va contribuer à l'amélioration des connaissances sur la biologie de ces champignons biotrophes et contribuera à faciliter la recherche de nouvelles méthodes de lutte contre la maladie / The leaf rust disease caused by Melampsora larici-populina (Mlp) is the main disease affecting poplar plantations in Europe with severe economic losses. The recent sequencing of the genome of Mlp (strain 98AG31) opens new perspectives to identify key genes involved in the fungal infection process and particularly those encoding fungal effectors that could manipulate host cell structure and function to facilitate host colonization. Analysis of the rust transcriptome during time course infection of poplar leaves, based on NimbleGen oligoarrays and massive EST sequencing led to the identification of genes related to fungal germination, biotrophy and sporulation. A consistant induction of genes encoding small-secreted proteins (SSPs) was observed during the biotrophic growth at 96 hours post-inoculation (hpi) but also at 168 hpi in the palisade mesophyll using laser capture microdissection. Mlp Secretome analysis, based on annotation, evolution and expression of genes encoding SSPs helped in identifying candidate poplar rust effectors. Some, specifically expressed in planta or showing homologies with known rust effectors were localized around the haustorium. Interestingly, other candidate genes, belonging to multigenic families under diversifying selection are cystein-rich, specifically expressed in planta and harbour a translocation signal potentially involved in effector export inside host cell. This genome-wide analysis of putative fungal effectors will contribute to the general knowledge of rust biology and will help to set new approaches to prevent and control the disease

En cancérologie, les puces à ADN mesurant le transcriptome sont devenues un outil commun pour chercher à caractériser plus finement les pathologies, dans l’espoir de trouver au travers des expressions géniques : des mécanismes,des classes, des associations entre molécules, des réseaux d’interactions cellulaires. Ces réseaux d’interactions sont très intéressants d’un point de vue biologique car ils concentrent un grand nombre de connaissances sur le fonctionnement cellulaire. Ce travail de thèse a pour but, à partir de ces mêmes données d’expression, d’extraire des structures pouvant s’apparenter à des réseaux d’interactions génétiques. Le cadre méthodologique choisi pour appréhender cette problématique est les « Réseaux Bayésiens », c’est-à-dire une méthode à la fois graphique et probabiliste permettant de modéliser des systèmes pourtant statiques (ici le réseau d’expression génétique) à l’aide d’indépendances conditionnelles sous forme d’un réseau. L’adaptation de cette méthode à des données dont la dimension des variables (ici l’expression des gènes, dont l’ordre de grandeur est 105) est très supérieure à la dimension des échantillons (ordre102 en cancérologie) pose des problèmes statistiques (de faux positifs et négatifs) et combinatoires (avec seulement 10gènes on a 4×1018 graphes orientés sans circuit possibles). A partir de plusieurs problématiques de cancers (leucémies et cancers du sein), ce projet propose une stratégie d’accélération de recherche de réseaux d’expression à l’aide de Réseaux Bayésiens, ainsi que des mises en œuvre de cette méthode pour classer des tumeurs, sélectionner un ensemble de gènes d’intérêt reliés à une condition biologique particulière, rechercher des réseaux locaux autour d’un gène d’intérêt.On propose parallèlement de modéliser un Réseau Bayésien à partir d’un réseau biologique connu, utile pour simuler des échantillons et tester des méthodes de reconstruction de graphes à partir de données contrôlées. / In oncology, microarrays have become a classical tool to search and characterize pathologies at a deeper level than previous methods, using genetic expression to find the mechanisms, classes, molecular associations, and cellular interaction networks of different cancers. From a biological point of view, these cellular networks are interesting because they concentrate a large amount of knowledge about cellular processes. The goal of this PhD thesis project is to extract structures that could correspond to genetic interaction networks from the expression data. "Bayesian Networks", i.e. a graphic and probabilistic method that models even static systems (like the expression network) with conditional independences, are used as the framework to investigate this problem. The adaptation of this method to data where the dimension of the variables (about 105 for gene expression) is much greater than the dimension of the samples (about 102 in oncology) aggravates some statistical and combinatorial problems. For several cancer problematics, this project proposes an acceleration strategy for capturing expression networks with Bayesian Networks and some methods to classify tumors, finding gene signatures of particular biological conditions by searching for local networks in the neighborhood of a gene of interest. In parallel, we propose to model a Bayesian Network from a known biological network, which is useful to simulate samples and to test these methods to reconstruct graphs from

Les béta-rhizobia sont des symbiotes de légumineuses retrouvées principalement associés au genre Mimosa. Les études des symbiotes de Mimosa pudica révèlent différents profils de diversité au sein des alpha (Rhizobium spp) et béta-rhizobia (Burkholderia, Cupriavidus) le long de la ceinture tropicale, les béta-rhizobia étaient toujours majoritaires dans les nodosités de cette plante hôte. Dans ce travail de thèse, nous avons étudié cette spécificité d'association béta-rhizobia/Mimosa pudica, par une approche couplant l'étude des traits symbiotiques bactériens à l'analyse des profils d'expression de leurs génomes dans les premières étapes de la symbiose, et en comparaison avec les alpha-rhizobia. Nous avons analysé les traits symbiotiques (compétitivité pour la nodulation, efficience symbiotique) au niveau intra et interspécifique de 4 espèces de béta-rhizobia et 4 d'alpha-rhizobia. Si l'efficience symbiotique est similaire parmi toutes les souches testées, différents niveaux de compétitivité ont été trouvés selon l'espèce, B. phymatum et B. tuberum étant les plus compétitives. Les tests effectués sur différentes variétés de M. pudica montrent un effet variétal sur la compétitivité de C. taiwanensis. Les traits symbiotiques mesurés expliquent en partie les profils de diversité des symbiotes de M. pudica dans les zones d'origine (Amérique du Sud) ou en zone introduite (Taiwan). Les transcriptomes de trois bactéries ayant des traits symbiotiques différents (B. phymatum STM815, C. taiwanensis LMG19424 et R. mesoamericanum STM3625) ont été comparés (par RNAseq), pour relier les différentes réponses induites par les exsudats racinaires aux traits symbiotiques de chaque rhizobium. Chaque bactérie développe une stratégie spécifique liée à ses traits symbiotiques et à l'origine de la symbiose dans son groupe bactérien. / Beta-rhizobia are legume symbionts mainly found associated to the Mimosa genus. Diversity studies of Mimosa pudica symbionts in native and introduced areas reveal different diversity patterns of alpha (Rhizobium spp) and beta-rhizobia (Burkholderia, Cupriavidus), with beta-rhizobia being always the main symbionts in the nodules of this legumes species. In this thesis we have studied the symbiotic specificity between beta-rhizobia and M. pudica (and the comparison with alpha-rhizobia) by a dual approach combining the study of bacterial symbiotic traits and the analysis of their transcriptomes in the first steps of symbiosis. We analysed symbiotic traits (nodulation competitiveness, symbiotic efficiency) at intra and interspecific levels on four species of beta-rhizobia and four of alpha-rhizobia. If symbiotic efficiency is similar among all strains, different levels of competitiveness were measured with a strong strain effect largely explained by the species affiliation, B. phymatum and B. tuberum being the most competitive species. Tests on different M. pudica varieties showed an impact on the competitiveness of C. taiwanensis. Symbiotic traits explained in part the symbiont patterns observed in diversity studies in French Guiana (M. pudica native area) and Taiwan (introduced). Root-exudates induced transcriptomes of three bacteria (two beta--rhizobia: B. phymatum STM815, C. taiwanensis LMG19424 and one alpha, R. mesoamericanum STM3625) with contrasted symbiotic traits were compared (by RNAseq). Each bacterium develops a specific strategy linked to its symbiotic traits and the origin of symbiosis in its bacterial group.

Les souches Saccharomyces cerevisiae présentent une forte diversité phénotypique, notamment au niveau des propriétés fermentaires parmi les souches œnologiques. Les bases moléculaires de ces différences de comportements ainsi que les mécanismes d'adaptation à l'environnement œnologique sont encore mal connues. Les variations d'expression génique contribuent à cette diversité phénotypique, cependant les réseaux et les gènes concernés sont peu décrits. Afin d'aborder ces questions, nous avons développé une approche intégrée « génétique-génomique » qui combine la cartographie de QTL en fermentation alcoolique et les profils d'expression d'une population recombinée. La recherche de QTL d'expression nous a permis de détecter 1465 eQTL correspondant à des régulations locales ou distantes dont certaines regroupées en hotspots. Nous avons déterminé qu'une duplication d'un segment chromosomique est impliquée dans le contrôle de la cinétique de fermentation. Nos données ont révélé que les perturbations d'expression concernent de nombreux réseaux métaboliques. Nous avons caractérisé plus finement les sources de variations d'expression qui affectent les systèmes de détoxification, et avons montré que les modifications de régulation de plusieurs exporteurs membranaires sont liées à des mutations dans des facteurs de transcription. Nous avons également décrypté les variations contrôlant les gènes du métabolisme de la thiamine. Nous montrons qu'une altération du senseur Thi3p, conduit à privilégier l'expression des gènes de biosynthèse de la thiamine chez la souche œnologique aux dépens d'un gène de pyruvate décarboxylase. Ces travaux nous ont ainsi permis d'accéder à une partie des bases moléculaires responsables de la diversité et de l'adaptation des souches aux conditions œnologiques. / Saccharomyces cerevisiae strains have a high phenotypic diversity, particularly in terms of fermentation properties among wine strains. The molecular bases underlying these behavior differences and the mechanisms of adaptation to the wine environment are still poorly known. Changes in gene expressions contribute to this phenotypic diversity. However, the regulatory networks and the genes involved are badly described. To address these questions, we developed an integrated “genetics-genomics” approach that combines QTL mapping in alcoholic fermentation and expression analysis of a recombinant population. The search for expression QTL allowed us to detect 1465 eQTL corresponding to local or distant regulations, several of them being grouped into hotspots. We highlighted that a duplication of a chromosomal segment is involved in the control of the fermentation kinetics. Our data showed that the expression disturbances are involved in many metabolic networks. We characterized more precisely the sources of expression variations affecting the detoxification systems, and showed that the changes in regulation of several membrane exporters are linked to mutations in transcription factors. We have also deciphered the variations controlling genes of the thiamine metabolism. We showed that an alteration of the sensor Thi3p tends to favor the expression of genes for the thiamine biosynthesis in wine strain at the expense of a gene encoding a pyruvate decarboxylase. This work allowed us to access some of the molecular bases responsible of the diversity and the strains adaptation to oenological conditions.

Les ARN longs non-codants (lncRNAs) constituent une famille d'ARN hétérogènes qui jouent un rôle majeur dans de nombreux cancers et notamment dans les mélanomes. Le chien est un modèle naturel et spontané pour l’analyse génétique comparée des cancers et, l'annotation du génome canin a récemment été enrichie avec l'identification de plus 10 000 lncRNAs. Afin de réaliser des prédictions fonctionnelles bioinformatiques des lncRNAs, nous avons caractérisé les profils d'expression des lncRNAs canins à partir de 26 tissus distincts. Nous avons défini la spécificité tissulaire de l’expression des lncRNAs et inféré leur fonctionnalité potentielle par des analyses de génomique et de transcriptomique comparatives avec des données humaines issues du projet ENCODE (ENCyclopedia Of DNA Elements). Comme chez l'homme et la souris, une grande proportion de lncRNAs canins (44 %) est exprimée de manière spécifique au sein d’un tissu. Par une approche de génomique comparative, nous avons identifié plus de 900 lncRNAs orthologues entre l’homme et le chien et pour 26 % d’entre eux, des patrons d'expression entre tissus significativement conservés (p < 0,05). Dans le cadre de l'étude des mélanomes canins, nous avons analysé les données de RNA-seq de 52 échantillons tumeurs/contrôles de mélanomes oraux. Nous avons identifié plus de 750 lncRNAs différentiellement exprimés entre la tumeur et le contrôle (FDR < 0,01), dont plus de 100 conservés avec l’homme. Ces lncRNAs constituent de bons candidats pour étudier la régulation de la progression tumorale des mélanomes chez le chien et pourront être évalués pour leurs potentiels diagnostic et thérapeutique en médecine humaine et vétérinaire. / Long non-coding RNAs (lncRNAs) are a family of heterogeneous RNAs that play a major role in many cancers, particularly in melanomas. The dog is a natural and spontaneous model for the comparative genetic analysis of cancers and, the annotation of the canine genome has recently been enriched with the identification of over 10,000 lncRNAs. In order to perform functional bioinformatic predictions of lncRNAs, we have characterized the expression patterns of canine lncRNAs from 26 distinct tissues representative of the major functions of the organism. We defined the tissue specificity of lncRNAs expression and inferred their potential functionality by comparative genomic and transcriptomic analyses with human data from the ENCODE project (ENCyclopedia Of DNA Elements). As in humans and mice, we show that a large proportion of canine lncRNAs (44%) are expressed specifically within a tissue. Using a comparative genomic approach, we have identified more than 900 orthologue lncRNAs between humans and dogs, and we show that for 26% of them, tissue expression patterns are also significantly conserved (p < 0.05). In the study of canine melanomas, we investigated the lncRNAs from RNA-seq data from 52 tumour/control samples of oral melanoma. We identified more than 750lncRNAs differentially expressed between tumour and control (FDR < 0.01), of which more than 100 were conserved with humans. These lncRNAs are good candidates to study the regulation of tumour progression of melanomas in dogs and can be evaluated for their diagnostic and therapeutic potential in human and veterinary medicine.