Dissertationen zum Thema „Analyse RNAseq“

L'épissage alternatif est un processus biologique qui génère la diversité du protéome malgré le nombre limité de gène. Ce mécanisme régule à la fois les gènes de manières qualitatives (isoformes exprimées) mais aussi quantitatives (niveau d'expression). Avec le développement des technologies de séquençage à haut débit, il est maintenant possible d'étudier à large échelle les aspects quantitatif et qualitatif du transcriptome avec une même expérience (RNA-seq). Durant ma thèse, j'ai développé une nouvelle méthode d'analyse de l'épissage alternatif dans les données RNA-seq. J'ai également participé à la mise en place du pipeline global d'analyse de données RNA-seq (expression et épissage) qui a été utilisé pour analyser un grand nombre de jeux de données. Dans un second temps, nous avons comparé notre outil d'analyse de l'épissage, FaRLine, qui est basé sur l'alignement sur un génome de référence, à KisSplice, une méthode basée sur l'assemblage. Nous avons montré que ces méthodes trouvaient un grand nombre d'événements en communs (70%), mais qu'il existait des différences non négligeables dues à la méthodologie. Nous avons analysé et classifié ces événements en 4 grandes catégories. Les variants faiblement exprimés et les exons chevauchant des éléments répétés sont mieux annotés par les méthodes basées sur l'alignement. Alors que les méthodes basées sur l'assemblage trouvent des nouveaux variants (exons ou sites d'épissage non annotés) et prédisent de l'épissage alternatif dans les famille de gènes paralogues. Notre travail souligne les points qui nécessitent encore l'amélioration des méthodes bioinformatiques. Enfin, j'ai participé au développement de méthodes permettant d'aider les biologistes à évaluer l'impact fonctionnel de modification d'épissage, que ce soit au niveau de la protéine produite (annotation des domaines protéiques au niveau des exons), ou à un niveau plus global en intégrant les modifications d'épissage dans les voies de signalisation
Alternative splicing is the biological process that explain the large diversity of the proteome compared to the limited number of genes. This process allow a qualitative regulation (expressed isoforms) and a quantitative regulation (expression level). The growth of high-trhoughtput sequencing methods enabled the analysis of these two aspects (quantitative and qualitative regulation) with the same experiment (RNA-Seq). During my PhD, I developped a new tool to analyse alternative splicing from RNA-Seq data. I also participated in the automatisation of the complet pipeline of RNA-Seq analysis (expression and splicing). This pipeline has been used to analyse various datasets. Then, we compared our mapping-first tool, FaRLine, with an assembly-first method, KisSplice. We found that the predictions of the two pipelines overlapped (70\% of exon skipping events were common), but with noticeable differences. The mapping-first approach allowed to find more lowly expressed splicing variants, and was better in predicting exons overlapping repeated elements. The assembly-first approach allowed to find more novel variants, including novel unannotated exons and splice sites. It also predicted AS in families of paralog genes. Our work point out where the bioinformatic improvment are still needed. Finally, I participated in the developpement of bioinformatics methods to help biologists to evualuate the fonctionnal impact of splicing alteration : at the level of the protein product by annotating fonctionnal domain at the exon level or at a more global level, by integrating splicing modifications in signaling pathways

Élucider les processus transcriptionnels et épigénétiques dérégulés dans les cancers est fondamental pour mieux comprendre les voies biologiques impliquées et proposer une thérapie adaptée au phénotype moléculaire de chaque tumeur. Les approches classiques de classification non supervisée définissent des groupes moléculaires principaux pour chaque type tumoral. Cependant, ces méthodes, appliquées à des tumeurs complexes comme le carcinome hépatocellulaire (CHC), le 3ème cancer le plus mortel au monde, définissent des groupes qui restent relativement hétérogènes et ne reflètent qu’imparfaitement la diversité des mécanismes biologiques à l’œuvre dans ces tumeurs. Au cours de ma thèse, j’ai développé une stratégie d’analyses innovante, basée sur l’analyse en composantes indépendantes (ACI), pour extraire des signatures de processus biologiques précis à partir de grands jeux de données transcriptomiques et épigénétiques. Grace à cette nouvelle approche, j’ai identifié des groupes de gènes co-régulés, associés à des phénotypes ou altérations moléculaires précises. De même, l’analyse en composantes indépendantes du méthylome de 738 CHC m’a permis d’isoler 13 signatures épigénétiques stables, préférentiellement actives dans certaines tumeurs et certains sites CpG. Ces signatures incluent des signatures de méthylation précédemment associées au vieillissement et au cancer, mais aussi de nouvelles signatures d'hyper- et d'hypométhylation liées à des événements « drivers » et sous-groupes moléculaires spécifiques. Ces résultats nous éclairent sur la diversité des mécanismes moléculaires impliqués dans la carcinogenèse hépatique. Les outils d’analyse biostatistique innovants que j’ai développés ont été incorporés dans un package R librement utilisable par la communauté scientifique
Elucidating deregulated transcriptional and epigenetic processes in cancers is fundamental to better understand the biological pathways involved and to propose a therapy adapted to the molecular phenotype of each tumor. Classical unsupervised classification approaches define, for each tumor type, the main molecular groups. However, these methods, applied to complex tumors such as hepatocellular carcinoma (HCC), the 3rd cause of cancer-associated mortality worldwide, define groups that remain relatively heterogeneous and only imperfectly reflect the diversity of biological mechanisms at work in these tumors. During my PhD, I developed a, innovative strategy involving independent component analysis (ICA) to extract signatures of precise biological processes in large transcriptomic and epigenetic tumor data sets. This new approach allowed me to identify groups of co-regulated genes associated with specific phenotypes or molecular alterations. Similarly, independent component analysis of the methylomes of 738 HCC revealed 13 stable epigenetic signatures preferentially active in specific tumors and CpG sites. These signatures include signatures previously associated with ageing and cancer, but also new hyper- and hypomethylation signatures related to specific driver events and molecular subgroups. The work presented in this thesis sheds light on the diversity of molecular processes remodeling liver cancer transcriptomes and methylomes, improve the understanding of the molecular mechanisms involved in hepatic carcinogenesis and provides a statistical framework to unravel the signatures of these processes

Le développement du séquençage ARN, ou RNAseq, a permis l'essor de la recherche intensive de biomarqueurs dans de nombreux domaines de la biologie. L’information complète du transcriptome contenue dans les données de sorties, permet à un bioinformaticien assidu de dépasser les connaissances actuelles et d’accéder, grâce à des pipelines informatiques avancés, à d’innombrables signatures d’intérêts inédites. Dans cette thèse nous mettons en avant que ces marqueurs potentiels, essentiellement explorés pour répondre à des problématiques clinique en conditions pathologiques, peuvent être utilisés pour affiner la caractérisation de types de cellules sans marqueurs strictement spécifiques. Nous nous sommes intéressés aux cellules souches mésenchymateuses (MSCs), un type de cellules souches adultes multipotentes, fortement utilisées en clinique mais ne possédant pas de marqueurs positifs strictement spécifiques.Notre étude se concentre sur la recherche des ARN longs non-codants non annotés. Ces ARNs, aussi nommés "lncRNA", constituent une classe émergente de transcrits encore peu explorée à ce jour. De plus, cette catégorie démontre une spécificité conditionnelle et tissulaire élevée. Nous avons élaboré un pipeline d’analyse RNAseq optimisé pour la reconstruction et la quantification de lncRNAs non annotés.En utilisant les données publiques de RNAseq, venant de différentes sources de MSCs et d'autres types de cellules, nous avons identifié de nouveaux lncRNA non annotés exprimés spécifiquement dans les MSCs.Nous avons développé pour ce projet Kmerator.jl, un outil qui permet de décomposer un transcrit en sous séquences spécifiques (k-mers) afin de chercher et quantifier plus rapidement la signature de nos candidats dans un grand nombre de données RNAseq. Kmerator a également été utilisé dans d'autres applications pour tester la qualité des données RNA-seq disponibles en accés public.Après validation de ces nouveaux biomarqueurs de MSCs par qPCR, nous avons eu recours à plusieurs outils informatiques pour prédire leurs fonctions potentielles. Enfin, nous avons analysé des données RNAseq « single-cell » pour aborder l’hétérogénéité d’expression au sein des populations MSCs
The development of RNA sequencing, or RNAseq, have opened the path of intensive biomarkers research in many areas of biology. The complete information of the transcriptome contained in the output data, allows a bioinformatician to surpass the current knowledge and to access, thanks to advanced computer pipelines, to signatures of new interest. In this thesis, we are showing that these potential markers, classically used in clinical and pathological conditions, can be used to characterize cell types without extensive markers profile. We have studied mesenchymal stem cells, a type of adult multipotent stem cells, strongly used in clinics but without strickly specific positive markers. Our study mainly focuses on the search for non-annotated, long non-coding RNAs. These RNAs, also called "lncRNA", constitute an emerging class of transcripts and are still lightly explored.In addition, this category presents a highly tissue-related specificity. We have developed an optimized RNAseq pipeline for the reconstruction and quantification of non-annotated lncRNAs.Using public data from RNAseq, coming from different sources of MSC and other cell types, we have identified new non-annotated lncRNAs clearly and specifically expressed in MSCs. to complete this project, we developed Kmerator.jl, a bioinformatical tool that allows to decompose a transcript in k-mer, and select specific sub-sequences, in order to search and quantify at a faster rate the signature of our candidates in a large number of RNAseq dataset. After validation of these new biomarkers of MSCs by qPCR, we used several computer tools to predict their potential functions. Finally, we analyzed single-cell RNAseq data to address the heterogeneity of expression within MSC populations

Le virus Nipah (NiV) est un paramyxovirus hautement pathogène pour les humains faisant partie de la liste prioritaire pour la recherche et le développement de l’OMS. Les chauves-souris Pteropus sont le réservoir naturel asymptomatique du NiV et nous nous sommes intéressés aux mécanismes leur permettant de contrôler l’infection. Pour cela, nous avons réalisé une analyse transcriptomique comparative entre des cellules de chauves-souris et des cellules humaines. Nous avons tout d’abord observé des profils immunitaires innés distincts à l’état basal. Les cellules de chauves-souris présentent des niveaux élevés de récepteurs aux pathogènes comme TLR-3 et TLR-8, pouvant expliquer une détection rapide du virus. De plus, la cinétique d’activation des gènes a montré une différence entre les deux espèces, avec une activation précoce chez les chauves-souris alors que plus tardive et de plus grande amplitude chez les humains. L’activation précoce de la voie NF-kB a été observée chez les chauves-souris et parmi ces facteurs, c-Rel faisait partie des gènes les plus exprimés. L’analyse fonctionnelle a révélé que les protéines c-Rel humaine et de chauve-souris induisent l’activation de la voie NF-kB et sont inhibées par la protéine virale non structurale NiV-W. Nous avons aussi montré la capacité du c-Rel de chauve-souris, contrairement au c-Rel humain, de moduler l’activité du promoteur de réponse aux IFN (ISRE) après stimulation par l’IFN. Cette étude suggère que la réponse rapide et transitoire des Pteropus pourrait favoriser une meilleure régulation des réponses pro-inflammatoires et contribuer à leur capacité de contrôler l’infection NiV. Étant donné que nous ne disposons ni de traitement ni de vaccin pour ce virus, le travail a aussi porté sur l’évaluation d’un inhibiteur de fusion agissant sur l’entrée du virus, et un essai vaccinal chez un modèle primate. Ce dernier, combinant le récepteur CD40 à l’ectodomaine de la protéine NiV-G a été validé in vivo, démontrant une protection complète chez les singes immunisés. L’ensemble de ces résultats ouvre de nouvelles perspectives vers des approches antivirales innovantes
Nipah virus (NiV) is a highly pathogenic paramyxovirus for humans, listed as a priority for research and development by the WHO. Pteropus bats are the natural asymptomatic reservoir of NiV and we investigated on the mechanisms allowing them to control the infection. For this, we conducted a comparative transcriptomic analysis between bat and human cells. We observed distinct immune profiles at the basal state. Bat cells show high levels of receptors like TLR-3 and TLR-8, which may explain the rapid viral detection. Additionally, the kinetics of gene expression resulted to be different among the two species, as we detected early gene activation in bats, while the response in humans was delayed. Early activation of the NF-kB pathway was observed in bats, and among these factors, c-Rel was one of the most expressed genes. Functional analysis revealed that both human and bat c-Rel proteins induce NF-kB pathway activation and are inhibited by the non-structural protein NiV-W. We also demonstrated the ability of bat c-Rel, unlike human c-Rel, to modulate IFN response promoter (ISRE) activity after IFN stimulation. This study suggests that the rapid and transient response of Pteropus may promote better regulation of pro-inflammatory responses and contribute to their ability to control NiV infection. Since no treatment or vaccine is available for this virus, the work also focused on evaluating a fusion inhibitor acting on virus entry and a vaccine. The latter, combining the CD40 receptor with the ectodomain of the G protein, was validated in vivo, demonstrating complete protection Nipah virus (NiV) is a highly pathogenic paramyxovirus for humans, listed as a priority for research and development by the WHO. Pteropus bats are the natural asymptomatic reservoir of NiV and we investigated on the mechanisms allowing them to control the infection. For this, we conducted a comparative transcriptomic analysis between bat and human cells. We observed distinct immune profiles at the basal state. Bat cells show high levels of receptors like TLR-3 and TLR-8, which may explain the rapid viral detection. Additionally, the kinetics of gene expression resulted to be different among the two species, as we detected early gene activation in bats, while the response in humans was delayed. Early activation of the NF-kB pathway was observed in bats, and among these factors, c-Rel was one of the most expressed genes. Functional analysis revealed that both human and bat c-Rel proteins induce NF-kB pathway activation and are inhibited by the non-structural protein NiV-W. We also demonstrated the ability of bat c-Rel, unlike human c-Rel, to modulate IFN response promoter (ISRE) activity after IFN stimulation. This study suggests that the rapid and transient response of Pteropus may promote better regulation of pro-inflammatory responses and contribute to their ability to control NiV infection. Since no treatment or vaccine is available for this virus, the work also focused on evaluating a fusion inhibitor acting on virus entry and a vaccine. The latter, combining the CD40 receptor with the ectodomain of the G protein, was validated in vivo, demonstrating complete protection in immunized monkeys. These results open new perspectives for innovative antiviral approaches

>Magister Scientiae - MSc
Lung cancer remains the leading cause of cancer deaths worldwide, with the majority of cases attributed to non-small cell lung carcinomas. At the time of diagnosis, a large percentage of patients present with advanced stage of disease, ultimately resulting in a poor prognosis. The identification circulatory markers, overexpressed by the tumour tissue, could facilitate the discovery of an early, specific, non-invasive diagnostic tool as well as improving prognosis and treatment protocols. The aim was to analyse gene expression data from both microarray and RNA sequencing platforms, using bioinformatics and statistical analysis tools. Enrichment analysis sought to identify genes, which were differentially expressed (p < 0.05, FC > 2) and had the potential to be secreted into the extracellular circulation, by using Gene Ontology terms of the Cellular Component. Results identified 1 657 statically significant genes between normal and early lung cancer tissue, with only 1 gene differentially expressed (DE) between the early and late stage disease. Following statistical analysis, 171 DE genes selected as potential early stage biomarkers. The overall sensitivity of RNAseq, in comparison to arrays enabled the identification of 57 potential serum markers. These genes of interest were all downregulated in the tumour tissue, and while they did not facilitate the discovery of an ideal diagnostic marker based on the set criteria in this study, their roles in disease initiation and progression require further analysis.

This thesis utilised two approaches, forward genetics and comparative transcriptomic analysis, to investigate the contrasting response to anthocyanin production observed in two distinct Nicotiana benthamiana ecotypes, LAB and QLD. The thesis is a step forward in utilising N. benthamiana as a candidate in forward genetics, currently limited due to its large complex genome and polyploid nature. The study utilised a cross-population between LAB and QLD to investigate the nature of inheritance of the contrasting parental phenotypes in its progeny. Additionally, expression profiles of anthocyanin biosynthesis genes were analysed via differential expression analysis and a novel method of mathematical manifold analysis

En Amérique Latine, les punaises hématophages Triatominae transmettent à l’homme le parasite Trypanosoma cruzi, responsable de la maladie de Chagas touchant actuellement 5 millions de personnes. Même si les programmes d’éradication chimique des vecteurs sont efficaces, la maladie persiste du fait de la recolonisation des habitations humaines par des vecteurs provenant d’habitats naturels. Ainsi, certaines espèces présentent une capacité d’adaptation aux anthroposystèmes (processus de domiciliation), alors que d’autres espèces apparentées ne l’ont pas. Comprendre cette capacité d’adaptation est crucial d’un point de vue épidémiologique afin de cibler les espèces présentant un risque pour l’homme. La capacité à s’adapter à un nouvel habitat pourrait être liée à l’évolution du répertoire de gènes du système chimiosensoriel, important pour la perception du milieu. Cette étude a porté sur le système chimiosensoriel des Triatominae dans le but de documenter le processus d’adaptation et donc de domiciliation des vecteurs. Des données transcriptomiques obtenues en séquençage à haut débit ont été utilisées pour annoter et répertorier les gènes chimiosensoriels ainsi que pour comparer leur expression au sein de punaises hématophages d’habitats différents. L’existence d’une relation entre les variations de ces gènes chez différentes espèces de Triatominae et leur capacité d’adaptation à un habitat a par la suite été évaluée. L’espèce T. brasiliensis en voie de domiciliation au Brésil et présentant à la fois des populations sylvatiques, péri-domiciliaires et domiciliaires, et différentes espèces du genre Rhodnius d’habitats variés, ont été étudiées, notamment les deux espèces sœurs, R. robustus, sylvatique en Amazonie et R. prolixus majoritairement domiciliée dans toute son aire de répartition. En l’absence de génomes de références suffisamment proches de T. brasiliensis et des 10 espèces de Rhodnius étudiées, leurs transcriptomes ont été assemblés de novo. Les transcriptomes des deux espèces R. prolixus et R. robustus ont été assemblés par alignement sur le génome de R. prolixus. Chez ces différentes espèces de Triatominae étudiées, l’analyse du répertoire des gènes chimiosensoriels codant les OBPs et CSPs (familles multigéniques) comparé à celui d’autres Paranéoptères a montré des expansions géniques pouvant refléter des processus adaptatifs. Par ailleurs, chez les différentes espèces du genre Rhodnius, il existe une corrélation positive entre le nombre de gènes codant les OBPs et la capacité de domiciliation, suggérant l’implication de cette famille de gènes dans l’adaptation au milieu anthropique. Les analyses d’expression différentielle concernant les différentes populations de T. brasiliensis et les espèces R. prolixus/R. robustus ont montré qu’un certain nombre de transcrits sont différentiellement exprimés selon l’environnement dans lequel ont évolué les punaises notamment des gènes chimiosensoriels (OBPs, CSPs) ainsi que des gènes impliqués dans le rythme circadien et le comportement de recherche alimentaire (Takeout), dans la réponse à des stress environnementaux comme des gènes de détoxification (P450, glutathione S-transférase), dans la résistance à des changements climatiques (Heat-shock protéines) et dans la protection du milieu extérieur (protéines cuticulaires). Ce travail a permis de mettre à la disposition de la communauté scientifique des outils performants pour l’étude du processus de domiciliation des vecteurs de la maladie de Chagas (transcriptome, répertoire de gènes). Il a également permis de révéler des gènes qui pourraient être impliqués dans l’adaptation et/ou la plasticité phénotypique en réponse à un changement d’habitat. La compréhension des bases moléculaires de l’adaptation des vecteurs aux habitations humaines ouvre des potentialités de développer des méthodes alternatives de lutte contre les vecteurs qui pourraient être basées sur une perturbation de la communication chimique
In Latin America, the bloodsucking bugs (Triatominae, Hemiptera, Reduviidae) are vectors of the parasite Trypanosoma cruzi, which causes Chagas disease. More than five million people are infected. Even if chemical control campaigns are effective against vectors, the disease persists due to the recolonization of human habitations by vectors from natural habitats. Some species have the capacity to adapt to anthroposystems (domiciliation process), while other related species do not. Understanding this capacity to adapt is crucial from an epidemiological perspective to target species at risk to humans. The capacity to adapt to a new habitat could be linked to changes in the repertoire of chemosensory system genes, particularly for odorant binding proteins (OBP) and chemosensory proteins (CSP), which are important proteins to detect various odor stimuli. This study is based on the chemosensory system of Triatominae to document the adaptation process and then the domiciliation of the vectors. Transcriptomic data obtained by high-throughput sequencing were used to annotate and list the chemosensory genes and also to compare their expression in bloodsucking bugs from different habitats. The relationship between changes in these genes in different Triatominae species and their ability to adapt to a new habitat was evaluated. The species T. brasiliensis, which is in the process of domiciliation in Brazil with sylvatic, peridomiciliary and domiciliary populations, and various species of the genus Rhodnius from diverse habitats were studied, especially the two sibling species R. robustus, sylvatic in the Amazonia and R. prolixus mostly domiciliary throughout its geographical range. In the absence of a reference genome for T. brasiliensis, a reference transcriptome via de novo assembly (data 454 and Illumina) was achieved. The reference transcriptomes for 10 Rhodnius species were also established using the de novo assembly method. A genome reference based method on R. prolixus was also used to assemble the transcriptome of the two species R. prolixus and R. robustus. In the different species of the Triatominae studied, the chemosensory gene repertoire showed a high diversity and genic expansions compared to that of others Paraneoptera, which could reflect adaptive process. Furthermore, a positive correlation was shown between the number of OBP genes in Rhodnius species and their domiciliation ability, suggesting that this gene family is involved in the adaptation to anthropogenic environment. The differential expression analyses on the T. brasiliensis populations and the R. prolixus / R. robustus species showed that some transcripts are differentially expressed according to the environment in which the bugs have evolved, especially the chemosensory genes (OBP, CSP) and also genes involved in the circadian rhythm and foraging behavior (Takeout), in the response to environmental stress such as detoxification genes (P450, glutathione S-transferase), in resistance to climatic changes (heat-shock proteins) and in protection from the external environment (cuticular proteins).This work has helped make available to the scientific community powerful tools for studying the process of domiciliation of Chagas disease vectors (transcriptome, gene repertoire). It also revealed genes that could be involved in the adaptation and/or phenotypic plasticity in response to a change in habitat. Understanding the molecular basis of vector adaptation to human dwellings opens the potential to develop new tools to control the disease vectors, for example by disrupting chemical communication

Dans la première partie j'ai analysé des jeux de données de RNA-seq de transcriptome de petits ARNs disponibles dans les bases de données publiques. J'y ai observé 2 points intrigants : - une grande partie des lectures (bien que courtes) ne peux pas être alignée sur le génome de référence sans discordance et cette fraction non-alignable est parfois majoritaire. - de nombreuses lectures ont des tailles autours de 15-18nt qui ne correspondent à aucun type de petits ARNs connues, cette fraction est également majoritaires dans certains cas. Ces expériences sont souvent conçues pour la détection des miRNAs et l'analyse bioinformatique de ces données passent toujours par un alignement sur le génome de référence ou sur des séquences connues pour donner des petits ARNs. J'ai donc simplement éliminé la contrainte d'alignement dans l'analyse de ces données et effectué un regroupement des lectures par similarité (à la manière des ESTs). Ce regroupement donne une vision différente des données dans laquelle la notion de position génomique n'est plus centrale et ouvre la possibilité d'y découvrir des phénomènes non-standard. La deuxième partie est tirée d'une collaboration avec le laboratoire U675 INSERM. J'ai fait l'analyse bioinformatique des gènes dérégulés par la répression par RNAi du gène REST dans une lignée de neuroblastome de souris (N18). Ce gène est un facteur de transcription qui réprime les gènes neuronaux dans les cellules non neuronales. Ce répertoire de gènes dérégulés est potentiellement constitué de gènes clefs dans la biologie des neurones
In first part of this thesis I have analysed small RNA-seq transcriptome data. I have noticed : - a large fraction of reads can't be aligned perfectly on reference genome - lot of reads are very short (15-18 nt) and don't match on previously known functionnal small RNAs. These experiments are designed for miRNA discovery and bioinformatics analysis of these data use alignments on genome or on known small RNA precursors sequences. I have eliminated the alignment and I have clustered these sequences. This clustering let me to observe these data with a new view in wich the genomic location is not central and open the gate to discover unconventional events. The second part is the analysis of deregulate genes by the silencing of the gene REST/NRSF in mouse N18 cell line. This gene is a transcription factor and it works as a repressor of neuronal genes in non neuronal cells. This deregulate genes repertoire potentially contains key genes in neuron biology. We found in this repertoire a network of genes centered on SWI/SNF complex including SMARCA2. This gene was associated to schizophrenia (SZ) in association studies and structural variation studies. In this network we found another genes associated to SZ. We show that these genes exhibit positive evolution in primate compare to rodents

Die Infektionskrankheit Malaria wird von einzelligen Parasiten der Gattung Plasmodium verursacht und stellt vor allem im südlichen Afrika eine große Herausforderung dar. Die Zellen der Parasiten enthalten zwei endosymbiontische Organellen, den Apicoplast und das Mitochondrium. Beide Organellen besitzen ein reduziertes Genom. Die Struktur des mitochondrialen Genoms ist ungewöhnlich. Mit nur 6 kb gehört es zu den kleinsten beschriebenen Genomen und enthält neben drei proteinkodierende Gene auch 34 kleine rRNA-Gene. Um das Genom zu exprimieren wird eine Vielzahl von kernkodierten Faktoren benötigt. Die Regulation der Expression, die Prozessierung der polycistronischen Primärtranskripte und die Regulation des RNA-Metabolismus des Mitochondriums ist jedoch weitestgehend unbekannt. In dieser Arbeit konnten kurze RNAs in den Mitochondrien von P. falciparum mittels Hochdurchsatzsequenzierung identifiziert werden. Solche RNA-Akkumulationen an Transkriptenden werden in den Organellen höherer Pflanzen durch PPR-Proteine (Pentatricopeptide repeat) verursacht. Um zu untersuchen, ob in P. falciparum PPR-ähnliche, helikale Proteine vorhanden sind, wurde genomweit nach Proteinen mit repetitiven, helikalen Elementen gesucht. Dabei konnte eine vorher unbekannte Proteinfamilie identifiziert werden, die aufgrund ihrer 37 Aminosäure langen Motive Heptatricopeptide repeat Proteine (HPR) genannt wurde. In P. berghei konnte für 7 HPR-Proteine eine mitochondriale Lokalisation betätigt werden. Außerdem zeigten Deletionsversuche, dass die meisten HPR-Proteine in den Blutstadien essentiell sind. In vitro RNA-Bindestudien konnte für ein rekombinantes HPR-Protein eine unspezifische Interaktion mit mitochondrialen Transkripten nachgewiesen werden, während keine Bindung an DNA erfolgt. Eine breite Suche in verschiedenen phylogenetischen Gruppen zeigte, dass HPR-Proteine in verschiedensten eurkaryotischen Taxa vorhanden sind, mithin früh in der Evolution der eukaryotischen Zelle entstanden sind.
Malaria is caused by a single celled parasite of the genus Plasmodium. Especially in Sub-Saharan Africa, -this disease is a huge challenge for the health system. The cells of the parasites contain two organelles of endosymbiotic origin, the apicoplast and the mitochondrion. Both organelles still contain a reduced genome. For the expression of the genome, the organelles depends on a large set of nuclear encoded proteins. The mitochondrial genome has a unique structure. With only 6 kb it is one of the smallest genomes discovered to date and it contains only three protein coding genes along with 34 small ribosomal genes. The regulation of expression, the processing of the polycistronic primary transcript and the regulation of the RNA metabolism in the mitochondria of Plasmodium remains largely unknown. Through high-throughput sequencing of cellular RNA, we discovered a population of small RNAs originating in the mitochondria of P. falciparum. Similar RNA accumulations can be detected in the organelles of higher plants and are caused by helical-hairpin repeat proteins like PPR proteins (pentatricopeptide repeat). To search for plant-like RNA binding proteins similar to PPR proteins we scanned the nuclear genome of P. falciparum for helical-hairpin repeat proteins. We found a novel protein family with repetitive helical elements of 37 amino acid length we termed heptatricopeptide repeat (HPR) proteins. In the rodent Malaria parasite P. berghei, the mitochondrial localization for 7 HPR-Proteins was verified. In knockout studies, we also showed that almost all HPR proteins are essential for blood stages of P. berghei. In RNA-binding assays, one recombinant HPR protein showed unspecific interaction with mitochondrial transcripts but not with DNA. By broadening the search, we discovered that HPR proteins are found in multiple eukaryotic taxa.

The detection of non-protein-coding RNA (ncRNA) genes in bacteria and their diverse regulatory mode of action moved the experimental and bio-computational analysis of ncRNAs into the focus of attention. Regulatory ncRNA transcripts are not translated to proteins but function directly on the RNA level. These typically small RNAs have been found to be involved in diverse processes such as (post-)transcriptional regulation and modification, translation, protein translocation, protein degradation and sequestration. Bacterial ncRNAs either arise from independent primary transcripts or their mature sequence is generated via processing from a precursor. Besides these autonomous transcripts, RNA regulators (e.g. riboswitches and RNA thermometers) also form chimera with protein-coding sequences. These structured regulatory elements are encoded within the messenger RNA and directly regulate the expression of their “host” gene. The quality and completeness of genome annotation is essential for all subsequent analyses. In contrast to protein-coding genes ncRNAs lack clear statistical signals on the sequence level. Thus, sophisticated tools have been developed to automatically identify ncRNA genes. Unfortunately, these tools are not part of generic genome annotation pipelines and therefore computational searches for known ncRNA genes are the starting point of each study. Moreover, prokaryotic genome annotation lacks essential features of protein-coding genes. Many known ncRNAs regulate translation via base-pairing to the 5’ UTR (untranslated region) of mRNA transcripts. Eukaryotic 5’ UTRs have been routinely annotated by sequencing of ESTs (expressed sequence tags) for more than a decade. Only recently, experimental setups have been developed to systematically identify these elements on a genome-wide scale in prokaryotes. The first part of this thesis, describes three experimental surveys of exploratory field studies to analyze transcript organization in pathogenic bacteria. To identify ncRNAs in Pseudomonas aeruginosa we used a combination of an experimental RNomics approach and ncRNA prediction. Besides already known ncRNAs we identified and validated the expression of six novel RNA genes. Global detection of transcripts by next generation RNA sequencing techniques unraveled an unexpectedly complex transcript organization in many bacteria. These ultra high-throughput methods give us the appealing opportunity to analyze the complete RNA output of any species at once. The development of the differential RNA sequencing (dRNA-seq) approach enabled us to analyze the primary transcriptome of Helicobacter pylori and Xanthomonas campestris. For the first time we generated a comprehensive and precise transcription start site (TSS) map for both species and provide a general framework for the analysis of dRNA-seq data. Focusing on computer-aided analysis we developed new tools to annotate TSS, detect small protein-coding genes and to infer homology of newly detected transcripts. We discovered hundreds of TSS in intergenic regions, upstream of protein-coding genes, within operons and antisense to annotated genes. Analysis of 5’ UTRs (spanning from the TSS to the start codon of the adjacent protein-coding gene) revealed an unexpected size diversity ranging from zero to several hundred nucleotides. We identified and validated the expression of about 60 and about 20 ncRNA candidates in Helicobacter and Xanthomonas, respectively. Among these ncRNA candidates we found several small protein-coding genes that have previously evaded annotation in both species. We showed that the combination of dRNA-seq and computational analysis is a powerful method to examine prokaryotic transcriptomes. Experimental setups are time consuming and often combined with huge costs. Another limitation of experimental approaches is that genes which are expressed in specific developmental stages or stress conditions are likely to be missed. Bioinformatic tools build an alternative to overcome such restraints. General approaches usually depend on comparative genomic data and evolutionary signatures are used to analyze the (non-)coding potential of multiple sequence alignments. In the second part of my thesis we present our major update of the widely used ncRNA gene finder RNAz and introduce RNAcode, an efficient tool to asses local protein-coding potential of genomic regions. RNAz has been successfully used to identify structured RNA elements in all domains of life. However, our own experience and the user feedback not only demonstrated the applicability of the RNAz approach, but also helped us to identify limitations of the current implementation. Using a much larger training set and a new classification model we significantly improved the prediction accuracy of RNAz. During transcriptome analysis we repeatedly identified small protein-coding genes that have not been annotated so far. Only a few of those genes are known to date and standard proteincoding gene finding tools suffer from the lack of training data. To avoid an excess of false positive predictions, gene finding software is usually run with an arbitrary cutoff of 40-50 amino acids and therefore misses the small sized protein-coding genes. We have implemented RNAcode which is optimized for emerging applications not covered by standard protein-coding gene annotation software. In addition to complementing classical protein gene annotation, a major field of application of RNAcode is the functional classification of transcribed regions. RNA sequencing analyses are likely to falsely report transcript fragments (e.g. mRNA degradation products) as non-coding. Hence, an evaluation of the protein-coding potential of these fragments is an essential task. RNAcode reports local regions of high coding potential instead of complete protein-coding genes. A training on known protein-coding sequences is not necessary and RNAcode can therefore be applied to any species. We showed this with our analysis of the Escherichia coli genome where the current annotation could be accurately reproduced. We furthermore identified novel small protein-coding genes with RNAcode in this extensively studied genome. Using transcriptome and proteome data we found compelling evidence that several of the identified candidates are bona fide proteins. In summary, this thesis clearly demonstrates that bioinformatic methods are mandatory to analyze the huge amount of transcriptome data and to identify novel (non-)coding RNA genes. With the major update of RNAz and the implementation of RNAcode we contributed to complete the repertoire of gene finding software which will help to unearth hidden treasures of the RNA World.

Cette étude vise à explorer les mécanismes de tolérance des plantes aux xénobiotiques organiques de la famille des HAPs (phénanthrène), à travers l’analyse de l’impact des évènements de spéciation par hybridation et duplication génomique (allopolyploïdie). Nous avons pour cela mené une approche comparative sur un modèle de spéciation allopolyploïde récente, constitué des espèces parentales hexaploïdes S. alterniflora et S. maritima, et de l’allopolyploïde S. anglica qui résulte de la duplication du génome de leur hybride F1 S. x townsendii. Une approche intégrative basée sur des analyses physiologiques et moléculaires nous a permis de montrer que chez Spartina l’hybridation et le doublement du génome augmentent la tolérance aux xénobiotiques. Le parent paternel S. maritima se montre particulièrement sensible au phénanthrène par rapport au parent maternel S. alterniflora. Différentes analyses transcriptomiques ont permis l’identification de novo de transcrits spécifiquement exprimés en condition de stress, et l’annotation des petits ARNs (miARNs, leurs gènes cibles, et siARNs) agissant en tant que régulateurs de l’expression des gènes et la régulation des éléments transposables. Les analyses d’expression différentielle en réponse au stress ont permis de générer un modèle de régulation (miARN/gènes cibles) en réponse aux HAPs, testé par validation fonctionnelle en système hétérologue chez Arabidopsis. Un travail exploratoire de profilage du microbiome de la rhizosphère des spartines exposées au phénanthrène a été réalisé pour préciser les mécanismes de dégradation des xénobiotiques dans l’environnement en vue d’une application dans les stratégies de remédiation verte
We explored mechanisms involved in tolerance to organic xenobiotics belonging to PAHs (phenanthrene), in the context of allopolyploid speciation (hybrid genome duplication). We developed a comparative approach, using a recent allopolyploidization model including the hexaploid parental species S. alterniflora and S. maritima, and the allopolyploid S. anglica, which resulted from genome doubling of the F1 hybrid S. x townsendii. Integrative approach based on physiological and molecular analyses highlights that hybridization and genome doubling enhance tolerance to xenobiotics in Spartina. The paternal parent S. maritima exhibits higher sensitivity compared to the maternal parent S. alterniflora. Various transcriptomic analyses were performed, to identify de novo stress responsive transcripts, and to annotate small RNAs (miRNAs, their target genes, and siRNAs) involved in gene expression and transposable element regulations. Differential expression analyses in response to stress allowed us to develop a putative miRNA regulatory network (miRNA/target genes) in response to PAH, functionally validated in Arabidopsis as heterologous system. An exploratory profiling of Spartina rhizosphere microbiome exposed to phenanthrene was also performed to characterize environmental degradation abilities, in the perspective of optimizing green remediation strategies

Bacterial endophytes are the organisms that live inside the plant for a full or a part of their life cycle. Endophytic bacteria have captured the interest of agriculture industry due to their plant beneficial properties, such as synthesis of phytohormones, solubilization of soil nutrients, and alleviation of biotic and abiotic stresses. Several studies have reported that stress tolerant endophytic bacteria can work with a similar performance as non-stressed conditions when inoculated to the plants under stressed conditions. Combination of abiotic stresses such as salinity, drought and low nitrogen stress can have additive or agonistic effects on bacterial and plant growth, and their interactions. However, very few studies have reported the impact of combined stress on endophytic bacterial assisted plant growth promotion. Therefore, understanding the underlying mechanisms of endophytic bacterial assisted plant’s tolerance abiotic stresses may provide the means of better exploiting the beneficial abilities of endophytic bacteria in agricultural production. Thus, the aim of this thesis was to study the stress tolerance mechanisms, beneficial characteristics, and plant growth promotion characteristics of endophytic bacteria under individual and combined abiotic stresses. Transcriptome analysis of endophytic bacteria revealed that tolerance mechanisms to deal with one kind of stress is different than concurrent stresses. Salinity and drought stress largely modulated the genes involved in flagellar assembly and membrane transport, showing reduced motility under stress conditions to preserve the energy. Additionally, bacterial endophyte that can fix nitrogen was studied with maize plant growth promotion under drought and low nitrogen stress conditions. The results suggested that diazotrophic bacterial endophyte can promote plant growth under moderate individual and combined stress conditions. Plant growth promoting endophytic bacteria can be utilized as an efficient tool to increase crop production under individual and concurrent abiotic stresses.

Almost every precursor mRNA (pre-mRNA) in a eukaryotic organism undergoes splicing, in some cases resulting in the formation of more than one splice variant, a process called alternative splicing. RNA-Seq provides a major opportunity to capture the state of the transcriptome, which includes the detection of alternative spicing events. Alternative splicing is a highly regulated process occurring in a complex machinery called the spliceosome. In this dissertation, I focus on identification of different splice variants and splicing factors that are produced during Arabidopsis and soybean embryo development. I developed several data analysis pipelines for the detection and the functional characterization of active splice variants and splicing factors that arise during embryo development. The main goal of this dissertation was to identify transcriptional changes associated with specific stages of embryo development and infer possible associations between known regulatory genes and their targets. We identified several instances of exon skipping and intron retention as products of alternative splicing. The coding potential of the splice variants were evaluated using CodeWise. I developed CodeWise, a weighted support vector machine classifier to assess the coding potential of novel transcripts with respect to RNA secondary structure free energy, conserved domains, and sequence properties. We also examined the effect of alternative splicing on the domain composition of resulting protein isoforms. The majority of splice variants pairs encode proteins with identical domains or similar domains with truncation and in less than 10% of the cases alternative splicing results in gain or loss of a conserved domain. I constructed several possible regulatory networks that occur at specific stages of embryo development. In addition, in order to gain a better understanding of splicing regulation, we developed the concept of co-splicing networks, as a group of transcripts containing common RNA-binding motifs, which are co-expressed with a specific splicing factor. For this purpose, I developed a multi-stage analysis pipeline to integrate the co-expression networks with de novo RNA binding motif discovery at inferred splice sites, resulting in the identification of specific splicing factors and the corresponding cis-regulatory sequences that cause the production of splice variants. This approach resulted in the development of several novel hypotheses about the regulation of minor and major splicing in developing Arabidopsis embryos. In summary, this dissertation provides a comprehensive view of splicing regulation in Arabidopsis and soybean embryo development using computational analysis.
Ph. D.

2'-5'-oligoadenylate synthetases (OAS) are interferon-stimulated proteins that act in the innate immune response to viral infection. Upon binding to viral double-stranded RNAs, OAS enzymes produce 2'-5'-linked oligoadenylates that stimulate RNase L and ultimately slow viral propagation. Studies have linked mutations in the OAS1 gene to increased susceptibility to West Nile virus (WNV) infection, highlighting the importance of the OAS1 enzyme. Here I report that the 5'-terminal region (5'-TR) of the WNV genome, comprising both the 5'-untranslated region (5'-UTR) and initial coding region, is capable of OAS1 activation in vitro. This region contains three RNA stem loops (SLI, SLII, and SLIII), whose relative contribution to OAS1 binding affinity and activation were investigated using electrophoretic mobility shift assays and enzyme kinetics experiments. Stem loop I (SLI) is dispensable for maximum OAS1 activation, as a construct containing only SLII and SLIII was capable of enzymatic activation. Mutations to the RNA binding site of OAS1 confirmed the specificity of the interaction. Solution conformations of both the 5'-TR RNA of WNV and OAS1 were then elucidated using small-angle x-ray scattering. I also report that the 3' terminal region (3'-TR) is able to mediate specific interaction with and activation of OAS1. Binding and kinetic experiments identified a specific stem loop within the 3'-TR that is sufficient for activation of the enzyme. The solution confirmation of the 3'-terminal region was determined by small angle X-ray scattering, and computational models suggest a conformationally restrained structure comprised of a helix and short stem loop. Structural investigation of the 3'-TR in complex with OAS1 is also presented. Finally, we show that genome cyclization by base pairing between the 5'- and 3'-TRs, a required step for replication, is not sufficient to protect WNV from OAS1 recognition. The purity, monodispersity and homogeneity of all samples subjected to SAXS analysis were evaluated using dynamic light scattering and/or analytical ultra-centrifuge. These data provide a framework for understanding recognition of the highly structured terminal regions of a flaviviral genome by an innate immune enzyme.
October 2015

Bacterial endophytes are the organisms that live inside the plant for a full or a part of their life cycle. Endophytic bacteria have captured the interest of agriculture industry due to their plant beneficial properties, such as synthesis of phytohormones, solubilization of soil nutrients, and alleviation of biotic and abiotic stresses. Several studies have reported that stress tolerant endophytic bacteria can work with a similar performance as non-stressed conditions when inoculated to the plants under stressed conditions. Combination of abiotic stresses such as salinity, drought and low nitrogen stress can have additive or agonistic effects on bacterial and plant growth, and their interactions. However, very few studies have reported the impact of combined stress on endophytic bacterial assisted plant growth promotion. Therefore, understanding the underlying mechanisms of endophytic bacterial assisted plant’s tolerance abiotic stresses may provide the means of better exploiting the beneficial abilities of endophytic bacteria in agricultural production. Thus, the aim of this thesis was to study the stress tolerance mechanisms, beneficial characteristics, and plant growth promotion characteristics of endophytic bacteria under individual and combined abiotic stresses. Transcriptome analysis of endophytic bacteria revealed that tolerance mechanisms to deal with one kind of stress is different than concurrent stresses. Salinity and drought stress largely modulated the genes involved in flagellar assembly and membrane transport, showing reduced motility under stress conditions to preserve the energy. Additionally, bacterial endophyte that can fix nitrogen was studied with maize plant growth promotion under drought and low nitrogen stress conditions. The results suggested that diazotrophic bacterial endophyte can promote plant growth under moderate individual and combined stress conditions. Plant growth promoting endophytic bacteria can be utilized as an efficient tool to increase crop production under individual and concurrent abiotic stresses.

Après un cycle de la production des protéines, les sous-unités des ribosomes terminés sont dissociés, afin de les rendre disponibles pour de nouveaux cycles de traduction. Si lors de la traduction le ribosome pause, il ne pourra pas terminer la traduction et être recyclé par la voie classique. Un mécanisme de recyclage alternatif a évolué pour dissocier de tels ribosomes arrêtés. Un complexe composé des facteurs Dom34 et Hbs1 induit leur dissociation. Ce complexe est aussi impliqué dans des voies de contrôle qualité qui ciblent des ARN qui causent des arrêts ribosomiques. Dans cette thèse, l'importance de plusieurs sites fonctionnels du complexe Dom34-Hbs1 pour ces voies contrôle qualité des ARNs est étudiée. De plus, la relation entre ces voies et leurs détails sont examiné. Finalement, un nouveau rôle de Dom34-Hbs1, en dissociant des ribosomes inactifs ce qui rend leurs sous-unités disponibles pour de nouveaux cycles de la traduction, est décrit
Protein production is a cyclic process that consists of four stages: initiation, elongation, termination and recycling. During recycling the subunits of terminated ribosomes are dissociated, to make them available for new rounds of translation. If ribosomes stall during translation, ribosomes cannot terminate properly and canonical recycling cannot occur. Cells have mechanisms to rescue these stalled ribosomes. A complex formed by the factors Dom34 and Hbs1 induces their dissociation. This compex in RNA quality control, targeting RNAs that cause ribosomal stalling. In this thesis the importance of several functional sites of the Dom34-Hbs1 complex for the degradation of these RNA sis investigated. Details of and therelationship between RNA quality control pathways in which the complex functions are further investigated. Finally, a new role of this complex, dissociating inactive ribosomes and there by making their subunits available to re-enter the translation cycle is described

Submitted by Silvana Teresinha Dornelles Studzinski (sstudzinski) on 2015-05-04T17:25:43Z No. of bitstreams: 1 Rodrigo Morais.pdf: 5083865 bytes, checksum: 69563cc7178422ac20ff08fe38ee97de (MD5)
Made available in DSpace on 2015-05-04T17:25:43Z (GMT). No. of bitstreams: 1 Rodrigo Morais.pdf: 5083865 bytes, checksum: 69563cc7178422ac20ff08fe38ee97de (MD5) Previous issue date: 2013
Nenhuma
A área de Mineração de Opiniões e Análise de Sentimentos surgiu da necessidade de processamento automatizado de informações textuais referentes a opiniões postadas na web. Como principal motivação está o constante crescimento do volume desse tipo de informação, proporcionado pelas tecnologia trazidas pela Web 2.0, que torna inviável o acompanhamento e análise dessas opiniões úteis tanto para usuários com pretensão de compra de novos produtos quanto para empresas para a identificação de demanda de mercado. Atualmente, a maioria dos estudos em Mineração de Opiniões e Análise de Sentimentos que fazem o uso de mineração de dados se voltam para o desenvolvimentos de técnicas que procuram uma melhor representação do conhecimento e acabam utilizando técnicas de classificação comumente aplicadas, não explorando outras que apresentam bons resultados em outros problemas. Sendo assim, este trabalho tem como objetivo uma investigação empírica e comparativa da aplicação do modelo clássico de Redes Neurais Artificiais (RNAs), o multilayer perceptron , no problema de Mineração de Opiniões e Análise de Sentimentos. Para isso, bases de dados de opiniões são definidas e técnicas de representação de conhecimento textual são aplicadas sobre essas objetivando uma igual representação dos textos para os classificadores através de unigramas. A partir dessa reresentação, os classificadores Support Vector Machines (SVM), Naïve Bayes (NB) e RNAs são aplicados considerandos três diferentes contextos de base de dados: (i) bases de dados balanceadas, (ii) bases com diferentes níveis de desbalanceamento e (iii) bases em que a técnica para o tratamento do desbalanceamento undersampling randômico é aplicada. A investigação do contexto desbalanceado e de outros originados dele se mostra relevante uma vez que bases de opiniões disponíveis na web normalmente apresentam mais opiniões positivas do que negativas. Para a avaliação dos classificadores são utilizadas métricas tanto para a mensuração de desempenho de classificação quanto para a de tempo de execução. Os resultados obtidos sobre o contexto balanceado indicam que as RNAs conseguem superar significativamente os resultados dos demais classificadores e, apesar de apresentarem um grande custo computacional para treinamento, proporcionam tempos de classificação significantemente inferiores aos do classificador que apresentou os resultados de classificação mais próximos aos dos resultados das RNAs. Já para o contexto desbalanceado, as RNAs se mostram sensíveis ao aumento de ruído na representação dos dados e ao aumento do desbalanceamento, se destacando nestes experimentos, o classificador NB. Com a aplicação de undersampling as RNAs conseguem ser equivalentes aos demais classificadores apresentando resultados competitivos. Porém, podem não ser o classificador mais adequado de se adotar nesse contexto quando considerados os tempos de treinamento e classificação, e também a diferença pouco expressiva de acerto de classificação.
The area of Opinion Mining and Sentiment Analysis emerges from the need for automated processing of textual information about reviews posted in the web. The main motivation of this area is the constant volume growth of such information, provided by the technologies brought by Web 2.0, that makes impossible the monitoring and analysis of these reviews that are useful for users, who desire to purchase new products, and for companies to identify market demand as well. Currently, the most studies of Opinion Mining and Sentiment Analysis that make use of data mining aims to the development of techniques that seek a better knowledge representation and using classification techniques commonly applied and they not explore others classifiers that work well in other problems. Thus, this work aims a comparative empirical research of the ap-plication of the classical model of Artificial Neural Networks (ANN), the multilayer perceptron, in the Opinion Mining and Sentiment Analysis problem. For this, reviews datasets are defined and techniques for textual knowledge representation applied to these aiming an equal texts rep-resentation for the classifiers. From this representation, the classifiers Support Vector Machines (SVM), Naïve Bayes (NB) and ANN are applied considering three data context: (i) balanced datasets, (ii) datasets with different unbalanced ratio and (iii) datasets with the application of random undersampling technique for the unbalanced handling. The unbalanced context inves-tigation and of others originated from it becomes relevant once datasets available in the web ordinarily contain more positive opinions than negative. For the classifiers evaluation, metrics both for the classification perform and for run time are used. The results obtained in the bal-anced context indicate that ANN outperformed significantly the others classifiers and, although it has a large computation cost for the training fase, the ANN classifier provides classification time (real-time) significantly less than the classifier that obtained the results closer than ANN. For the unbalanced context, the ANN are sensitive to the growth of noise representation and the unbalanced growth while the NB classifier stood out. With the undersampling application, the ANN classifier is equivalent to the others classifiers attaining competitive results. However, it can not be the most appropriate classifier to this context when the training and classification time and its little advantage of classification accuracy are considered.

Over the last two decades the advancement in DNA sequencing technologies has enormously increased the amount of sequencing data available to researchers and geneticists. This has been accompanied by the development of tools for sequencing data analysis, including the human reference genome, that is undoubtedly an indispensable resource. It is known that the reference genome does not always represent the real consensus sequence of the human population, due to the inclusion of rare alleles and sequencing errors. Moreover, genomic duplications are often misassembled and, as a result, they may be found in the reference genome as a collapsed consensus, thus generating false variants. In this work I performed a thorough search for conflicting information between the human reference genome (GRCh37 and GRCh38) and some of the most popular human genetic resources such as the 1000 Genomes Project, to disclose minor alleles and to mine genetic inconsistencies. To search for unreported genomic duplications, I performed a genome wide screening for unbalanced heterozygosity. I found that inaccuracies and errors are much higher than expected. Minor alleles occurring with a frequency <10% are found on average every ~7,000 bases and include many rare variants that are never found elsewhere, producing high numbers of false positives as well as possible false negatives. The systematic screening for unbalanced heterozygosity revealed ~86,000 variants that are likely the result of unreported genomic duplications, involving functionally relevant genes such as MAP2K3 and KCNJ12. My findings may help the ongoing quest to obtain a highly accurate human genome reference sequence. Moreover, the results presented in this thesis will be useful to human geneticists in the process of filtering and selecting causative variants. The advancement in DNA sequencing technologies also accounts for the increasing usage of Whole Genome Sequencing approaches both in the research and clinical fields, thus revealing that the large majority of disease-associated SNPs are located in non-coding regions of the human genome. However, the functional interpretation of non-coding variants is still challenging. Part of my work also addressed this problem, aiming to develop a method for non-coding variant prioritization. The method, presented in the last chapter of this thesis, is based on a comparative genomics approach for the identification of functional constraints in primate orthologous genes. The first steps of my approach have proved to be powerful in identifying orthologous genes, but further work is necessary to optimize the multiple sequence alignment step and the identification of conserved domains.
Nel corso dell’ultimo ventennio l’avanzamento tecnologico nel campo del sequenziamento del DNA ha portato a un enorme aumento della quantità di dati di sequenziamento accessibili a ricercatori e genetisti. Questa crescita è stata accompagnata dallo sviluppo di strumenti necessari all’analisi dei dati; tra questi il genoma umano di riferimento è senza dubbio una risorsa indispensabile. È noto che il genoma di riferimento non sempre rappresenta la reale sequenza consenso della popolazione umana, poiché alleli rari ed errori di sequenziamento sono stati inclusi in essa. Inoltre, duplicazioni genomiche sono spesso mal assemblate e, di conseguenza, possono essere trovate nel genoma di riferimento come collassate, generando così false varianti. In questa tesi è descritta la ricerca approfondita di incongruenze tra il genoma umano di riferimento (GRCh37 e GRCh38) e alcune delle più popolari risorse di genetica umana, come il 1000 Genomes Project, per scovare alleli minori e inconsistenze genetiche. Per identificare duplicazioni genomiche non riportate nel genoma, è stata poi condotta un’ampia ricerca di eterozigosità sbilanciata. Questa analisi ha dimostrato che incongruenze ed errori sono molto più frequenti di quanto atteso. Infatti, alleli minori con una frequenza <10% sono stati trovati in media ogni ~7,000 basi e tra essi sono presenti molte varianti rare mai riportate nei database. Lo screening sistematico per l’eterozigosità sbilanciata ha mostrato inoltre che ~86,000 varianti possono derivare da duplicazioni genomiche non riportate nella sequenza di riferimento e che alcune di esse coinvolgono geni importanti come MAP2K3 e KCNJ12. I risultati descritti in questo lavoro possono contribuire alla definizione di una sequenza di riferimento del genoma umano altamente accurata. Inoltre, questi stessi risultati potranno essere utili ai genetisti umani nel processo di filtraggio e selezione delle varianti potenzialmente associate a malattie. L’avanzamento nel settore del sequenziamento del DNA ha condotto inoltre dell’utilizzo sempre maggiore degli approcci di sequenziamento dell’intero genoma, sia nel campo della ricerca sia nella diagnosi clinica, rivelando così che la gran parte degli SNP associati a malattia è localizzata nelle regioni non codificanti del genoma umano. Tuttavia, l’interpretazione funzionale delle varianti non codificanti è ancora una questione problematica. Parte del mio lavoro ha riguardato anche questo aspetto, con lo scopo di sviluppare un metodo per la prioritizzazione delle varianti non codificanti. Questo metodo, descritto nell’ultimo capitolo della tesi, si basa su un approccio di genomica comparata per l’identificazione di domini funzionali in geni ortologhi di organismi primati. I primi passaggi di questo approccio hanno dimostrato essere molto buoni per l’identificazione dei geni ortologhi, ma ulteriore lavoro è necessario per ottimizzare il processo di allineamento multiplo delle sequenze e l’identificazione dei domini conservati.

La maladie de Chagas est une maladie parasitaire causée par le protozoaire Trypanosoma cruzi et transmise par des insectes hématophages . Elle est composée de 2 phases : la phase aiguë et la phase chronique. Parmi les individus infectés, 30 % développent la forme chronique de la maladie. Les patients présentent des atteintes cardiaques, digestives (œsophage, côlon) et cardiodigestives. Notre étude a été focalisée sur les patients atteints de cardiomyopathie chagasique (CCC). Notre objectif est d’identifier des gènes de susceptibilité pouvant être impliqués dans le développement des formes chroniques. Notre étude a permis de mettre en évidence une variation d’expression de certains gènes entre les CCC et les contrôles. Nous nous sommes également intéressés aux processus épigénétiques pouvant réguler l’expression des gènes. Une étude de la méthylation de l’ADN croisée avec l’étude du transcriptome nous ont permis d’identifier des gènes présentant à la fois des variations d’expression et de méthylation. Pour certains de ces gènes, nous avons démontré que la méthylation est responsable de la variation d’expression observée. Enfin, nous avons étudié un ARN long non-codant, MIAT. Nous avons démontré qu’il est surexprimé chez les CCC par rapport aux contrôles et dans un modèle murin infecté par T. cruzi. De plus, l’analyse de l’expression de micro-ARNs couplée à une analyse de transcriptome nous a permis d'identifier plusieurs micro-ARNs indispensables à la régulation de l’expression des gènes. Enfin, une étude protéomique nous a permis de mettre en évidence une augmentation de la production de protéine pour certains gènes, en lien avec l’augmentation de l’expression observée
Chagas disease is a parasitic disease caused by the protozoan Trypanosoma cruzi and transmitted by the hematophagous insects. The disease is composed by acute and chronic phases. Among the infected individuals, 30 % develop chronic form. They suffer from heart, digestive (esophagus, colon) and cardiodigestives injury. Our study was focused on patients with dilated chagasic cardiomyopathy (CCC). Our goal is to identify susceptibility genes that may be involved in the development of chronic forms. Our study revealed a variation in the expression of certain genes between CCC group and controls. We are also interested in epigenetic processes that can regulate the expression of genes. A study of the DNA methylation crossed with the transcriptome allowed us to identify genes presenting both variations in expression and methylation. For some of these genes we demonstrated that methylation is responsible for the expression variation observed. Finally, we studied a long non-coding RNA called MIAT. Our study demonstrated that it is overexpressed in CCC compared to controls and in a murine model infected by T. cruzi. Furthermore, the analysis of the expression of micro-RNAs crossed with transcriptome analysis allowed us to identify several micro-RNAs whose functions are essential in the regulation of gene expression. Finally, a proteomic study allowed us to demonstrate an increase in the production of protein for certain genes, correlated with the increase in expression levels observed

Neurotrophins and their receptors are key molecules in the development of the
nervous system. Neurotrophin-3 binds preferentially to its high-affinity receptor
NTRK3, which exists in two major isoforms in humans, the full-length kinaseactive
form (150 kDa) and a truncated non-catalytic form (50 kDa). The two
variants show different 3'UTR regions, indicating that they might be differentially
regulated at the post-transcriptional level. In this work we explore how
microRNAs take part in the regulation of full-length and truncated NTRK3,
demonstrating that the two isoforms are targeted by different sets of microRNAs.
We analyze the physiological consequences of the overexpression of some of the
regulating microRNAs in human neuroblastoma cells. Finally, we provide
preliminary evidence for a possible involvement of miR-124 - a microRNA with no
putative target site in either NTRK3 isoform - in the control of the alternative
spicing of NTRK3 through the downregulation of the splicing repressor PTBP1.
Las neurotrofinas y sus receptores constituyen una familia de factores cruciales
para el desarrollo del sistema nervioso. La neurotrofina 3 ejerce su función
principalmente a través de una unión de gran afinidad al receptor NTRK3, del cual
se conocen dos isoformas principales, una larga de 150KDa con actividad de tipo
tirosina kinasa y una truncada de 50KDa sin dicha actividad. Estas dos isoformas
no comparten la misma región 3'UTR, lo que sugiere la existencia de una
regulación postranscripcional diferente. En el presente trabajo se ha explorado
como los microRNAs intervienen en la regulación de NTRK3, demostrando que las
dos isoformas son reguladas por diferentes miRNAs. Se han analizado las
consecuencias fisiológicas de la sobrexpresión de dichos microRNAs utilizando
células de neuroblastoma. Finalmente, se ha estudiado la posible implicación del
microRNA miR-124 en el control del splicing alternativo de NTRK3 a través de la
regulación de represor de splicing PTBP1.

High-throughput sequencing is a powerful tool to study diverse aspects of biology and applies to genome, transcriptome, and small RNA profiling. Ever increasing sequencing throughput and more specialized sequencing assays demand more sophisticated bioinformatics approaches. In this thesis, I present 4 studies for which I developed computational methods to handle high-throughput sequencing data to gain insights into biology. The first study describes the genome of High Five (Hi5) cells, originally derived from Trichoplusia ni eggs. The chromosome-level assembly (scaffold N50 = 14.2 Mb) contains 14,037 predicted protein-coding genes. Examination and curation of multiple gene families, pathways, and small RNA-producing loci reveal species- and order-specific features. The availability of the genome sequence, together with genome editing and single-cell cloning protocols, enables Hi5 cells as a new tool for studying small RNAs. The second study focuses on just one type of piRNAs that are produced at the pachytene stage of mammalian spermatogenesis. Despite their abundance, pachytene piRNAs are poorly understood. I find that pachytene piRNAs cleave transcripts of protein-coding genes and further target transcripts from other pachytene piRNA loci. Subsequently, systematic investigation of piRNA targeting by integrating different types of sequencing data uncovers the piRNA targeting rule. The third study describes computational procedures to map splicing branchpoints using high-throughput sequencing data. Screening >1.2 trillion RNA-seq reads determines >140,000 BPs for both human and mouse. Such branchpoints are compiled into BPDB (BranchPoint DataBase) to provide a comprehensive branchpoint catalog. The final study combines novel experimental and computational procedures to handle PCR duplicates that are prevalent in high-throughput sequencing data. Incorporation of unique molecular identifiers (UMIs) to tag each read enables unambiguous identification of PCR duplicates. Both simulated and experimental datasets demonstrate that UMI incorporation increases the reproducibility of RNA-seq and small RNA-seq. Surveying 7 common variables in high-throughput sequencing reveals that the amount of starting material and sequencing depth, but not the number of PCR cycles, determine the PCR duplicate frequency. Finally, I show that removing PCR duplicates without UMIs leads to substantial bias into data analysis.
2020-12-11T00:00:00Z

Η ριβονουκλεάση P (RNase P) είναι το ένζυμο που ευθύνεται για την ωρίμανση του 5΄ άκρου των πρόδρομων μορίων tRNA συμμετέχοντας έτσι στην πρωτεΐνοσύνθεση. H RNase P καταλύει την ενδονουκλεολυτική θραύση ενός φωσφοδιεστερικού δεσμού παρουσία ιόντων Mg2+ παράγοντας προϊόντα με 3΄ υδροξυλικά και 5΄ φωσφορικά άκρα. Η πλειοψηφία των RNase P ενζύμων που έχουν μελετηθεί μέχρι σήμερα είναι ριβονουκλεοπρωτεΐνες αποτελούμενες από μια RNA και πρωτεϊνικές υπομονάδες. Η RNA υπομονάδα της βακτηριακής RNase P είναι ένα από τα πρώτα καταλυτικά RNA που μελετήθηκαν. Η RNase P και το ριβόσωμα είναι τα μόνα γνωστά ριβοένζυμα που είναι συντηρημένα και στις τρείς φυλογεννετικές περιοχές. Δεδομένης της ζωτικής σημασίας των λεμφοκυττάρων στην ακεραιότητα του ανοσολογικού συστήματος και στην παθογένεια ευρέος φάσματος δερματολογικών και μη νοσημάτων, σκοπός της παρούσας εργασίας ήταν η ανάπτυξη μεθοδολογίας για την απομόνωση και τον καθαρισμό της RNase P από περιφερικά λεμφοκύτταρα υγιών μαρτύρων, ο προσδιορισμός της ενζυμικής της δραστικότητας, καθώς και η μελέτη της in vitro επιδράσεως δυο συνθετικών ρετινοειδών, του 13-cis ρετινοϊκού οξέος και της ασιτρετίνης, της αμινογλυκοσίδης νεομυκίνης Β, όπως και της διφωσφορικής χλωροκίνης στην δραστικότητα της λεμφοκυτταρικής RNase P. Το ένζυμο καθαρίστηκε με τη χρήση κατιοντοανταλλακτικής χρωματογραφίας φωσφοκυτταρίνης και προσδιορίστηκαν οι βέλτιστες συνθήκες δραστικότητάς του.Μετά από μελέτη της επίδρασης των συνθετικών ρετινοϊδών και της νεομυκίνης Β στην δραστικότητα της RNase P, προέκυψε ότι και τα τρία προκαλούν μια δοσοεξαρτώμενη αναστολή της δραστικότητας της RNase Ρ των ανθρωπίνων λεμφοκυττάρων. Η διφωσφορική χλωροκίνη δεν επηρεάζει τη δραστικότητα της λεμφοκυτταρικής RNase P. Λεπτομερής κινητική ανάλυση έδειξε πως η αναστολή που προκαλείται από την ασιτρετίνη είναι συναγωνιστικού τύπου, ενώ αυτή από την νεομυκίνη Β είναι μη συναγωνιστικού τύπου.Η κινητική σταθερά Km της λεμφοκυτταρικής RNase P βρέθηκε ότι ισούται με 245 nM και η Vmax με 0.42 pmol/min. Συμπερασματικά, η απομόνωση της RNase P από ανθρώπινα περιφερικά λεμφοκύτταρα καθιστά δυνατή τη μελέτη της πιθανής ανάμιξης αυτού του ριβοενζύμου στους παθογενετικούς μηχανισμούς διαφόρων αυτοάνοσων, φλεγμονωδών και νεοπλασματικών δερματοπαθειών και μπορεί να διευκολύνει τη περαιτέρω ανάπτυξη της βασιζομένης στην RNase P τεχνολογίας για τη γονιδιακή θεραπεία δερματικών και μη παθήσεων.
Ribonuclease P (RNase P) is the enzyme responsible for the 5΄ maturation of the precursor tRNA molecules, participating in tRNA biogenesis and therefore in protein synthesis. It catalyses the endonucleolytic cleavage of a phosphodiester bond in the presence of Mg2+ and results in the production of molecules that bear 3΄ hydroxyl and 5΄ phosphoric ends. Most forms of RNase P are ribonucleoproteins consisting of an essential RNA and protein subunits. The RNA component of the bacterial RNase P was one of the first identified catalytic RNAs. So far, RNase P and the ribosome are the only ribozymes known to be conserved in all kingdoms of life (bacteria, archaea and eucarya). In view of the vital importance of lymphocytes for an effective immune system, we proceeded to the RNase P isolation from human peripheral lymphocytes. The enzyme was purified with cation exchange phosphocellulose chromatography and the optimal conditions were determined. Herein, it was investigated the effect of the synthetic retinoids (cis-retinoic acid and acitretin), neomycin B, as well as chloroquine diphosphate, on the RNase P activity. Cis-retinoic acid, acitretin and neomycin B exerted a dose-dependent inhibitory effect on RNase P activity from human lymphocytes, wlile the activity was not affected in the presence of chloroquine diphosphate. A detailed kinetic analysis showed that the inhibition caused by acitretin was of competitive type, whereas that caused by neomycin B was of noncompetitive type. The kinetic constant Km of RNase P activity isolated from lymphocytes for the tRNA maturation reaction has been estimated equal to 245 nM and the Vmax value has been estimated equal to 0.42 pmol/min. Finally, the isolation of RNase P from human peripheral lymphocytes will enable the study of the possible involvement of this ribozyme in the pathogenetic mechanisms of diverse autoimmune, inflammatory and neoplastic cutaneous disorders and may facilitate the further development of RNase P-based technology for gene therapy of infectious and neoplastic dermatoses.

Background Within eukaryotes there is a complex cascade of RNA-based macromolecules that process other RNA molecules, especially mRNA, tRNA and rRNA. A simple example is the RNase MRP processing of ribosomal RNA (rRNA) in ribosome biogenesis. One hypothesis is that this complexity was present early in eukaryotic evolution; an alternative is that an initial simplified network later gained complexity by gene duplication in lineages that led to animals, fungi and plants. Recently there has been a rapid increase in support for the complexity-early theory because the vast majority of these RNA-processing reactions are found throughout eukaryotes, and thus were likely to be present in the last common ancestor of living eukaryotes, named here as the Eukaryotic Ancestor. Results We present an overview of the RNA processing cascade in the Eukaryotic Ancestor and investigate in particular, RNase MRP which was previously thought to have evolved later in eukaryotes due to its apparent limited distribution in fungi and animals and plants. Recent publications, as well as our own genomic searches have uncovered previously unknown RNase MRP RNAs, indicating that RNase MRP has a wide distribution in eukaryotes. Combining secondary structure and promoter region analysis of new and previously discovered RNase MRP RNAs along with analysis of the primary substrate (rRNA), allows us to discuss this distribution in the light of eukaryotic evolution. Conclusions We conclude that RNase MRP can now be placed in the RNA-processing cascade present in the Eukaryotic Ancestor. This highlights the complexity of RNAprocessing in early eukaryotes.

Dissertação de mestrado em Bioinformatics
In the past 30 years, accumulated evidence has been supporting viral infection as one factor responsible for 15-20% of human malignancies worldwide (W. S. Liang et al. 2014; McLaughlin-Drubin and Munger 2008). Studies on oncogenic viruses have proved their importance on cellular malfunction along the carcinogenic process, and showed that their association with cancer can amount from 15% to 100% (McLaughlin-Drubin and Munger 2008), depending on the type of tumour. With the large amount of genomic and metagenomic information available on public international consortia, such as TCGA database, it is nowadays possible to indirectly infer viral infections from the human centred omics studies, as a portion of the reads will align in viruses and bacteria. Taking as starting point the research made by Tang et al. 2013, we focused on cervical (CESC), hepatocellular (LIHC) and head and neck squamous cell (HNSC) carcinomas, which are known to show a high proportion of viral-positive cases (Tang et al. 2013). We downloaded RNAseq data from 309, 424 and 566 samples, respectively, and run the unmapped reads against a reference database of viruses (downloaded from NCBI) by using the tools Batch, SAMTOOLS, Bowtie and PRINTSEQ. Quantification of each virus was performed using parts per million reads (ppm) and only viruses with ppm above 10 were considered as positively infecting the sample. We confirmed that around 94% of CESC samples were infected, mostly by HPV (Human papillomavirus) and specifically by the HPV16 strain. Nearly 32% of LIHC were infected by HBV (hepatitis B virus). Almost 17% of HNSC samples were infected, and the HPV16 was the most common present virus. The evaluation of differential enrichment of metabolic pathways between infected and noninfected groups, for each cancer type, was performed in GSEA. Signs of enrichment for infection and immune related pathways were evident in CESC infected group, while in LIHC and HNSC infected groups the enrichment was mostly related with DNA replication and repair. This seems to indicate that infection is especially active in CESC, contradicting previous claims that tumorigenesis in cervix was not directly linked with infection. For the three cancer types, the viruses integrate their genome in the host genome, affecting DNA replication, maintenance and repair. In our investigation of integration of HPV16 genome in one HNSC tumor sample, we confirmed integration in the human RAD51B gene that codes a protein involved in DNA repair by homologous recombination. We thus confirmed that HPV16 can act both as indirect and direct carcinogen. The infection, most probably through the integration of the viral genome in the host genome, increased the amount of somatic mutations in the infected group in LIHC, but not in HNSC where tobacco consumption is also an important carcinogen. The low number of non-infected samples in CESC did not allow a reliable evaluation of changes in the amount of somatic mutations. Even so, in both LIHC and HNSC infected groups, some somatic mutations occurred in the context of immune-related pathways, showing that they can contribute to render these individuals susceptible to infection. Also, when checking expression of HPV16 genes in five samples each from CESC and HNSC, we confirmed that E6 and E7 genes are amongst the ones more expressed in many samples, while E2 is not expressed. E6 and E7 have been said to be preferentially integrated in the host genome, while E2, which controls their expression, is not integrated or it is disrupted. It is believed that the overexpression of E6 and E7 initiates carcinogenesis. The viral infection rates inferred here from mining the omics databases are very similar to the ones evaluated by standard methods (Tang et al. 2013), showing that public international consortia can indirectly provide interesting insights into the involvement of viral infection in tumorigenesis. The high number of samples per tumor, the wide geographic origin of the samples, and the high-throughput characterisation for different omics platforms allows multilayer comparisons and evaluations, in a scale not affordable before.
Nos últimos 30 anos foram-se acumulando evidências que têm vindo a apoiar a infecção viral como um factor responsável por 15-20% dos tumores malignos em humanos a nível mundial (W. S. Liang et al. 2014; McLaughlin-Drubin and Munger 2008). Estudos sobre os vírus oncogénicos demonstraram a sua importância no mau funcionamento celular ao longo do processo carcinogénico e demonstraram que a sua associação com o cancro varia entre 15% e 100% (McLaughlin-Drubin and Munger 2008), dependendo do tipo de tumor. Com a grande quantidade de informação genómica e metagenómica acessível nos consórcios internacionais públicos, tais como a base de dados TCGA, hoje em dia é possível inferir indiretamente infecções virais a partir de estudos genómicos centrados em humanos, uma vez que parte das reads irá alinhar com vírus e bactérias. Tomando como ponto de partida a pesquisa feita por Tang et al. 2013, concentramo-nos nos cancros cervical (CESC), hepatocelular (LIHC) e da cabeça e pescoço (HNSC), que são conhecidos por apresentar uma alta proporção de casos virais-positivos (Tang et al. 2013). Fizemos download de dados RNA-Seq de 309, 424 e 566 amostras, respectivamente, e comparamos unmapped reads contra uma base de dados viral de referência (retirada da base de dados do NCBI) usando as ferramentas Batch, SAMTOOLS, bowtie e PRINTSEQ. A quantificação de cada vírus foi feita usando partes por milhão (ppm) e apenas vírus com ppm acima de 10 foram considerados como estando a infectar positivamente uma amostra. Confirmamos que cerca de 94% das amostras de CESC foram infectadas, principalmente por HPV (papilomavírus humano) e, especificamente, pela estirpe HPV16. Quase 32% das amostras LIHC foram infectadas por HBV (vírus da hepatite B). E por volta de 17% de amostras HNSC foram infectadas e o HPV16 foi o vírus mais comum. A avaliação de enriquecimento diferencial de vias metabólicas entre grupos infectados e não infectados, para cada tipo de cancro, foi realizada por GSEA. Os sinais de enriquecimento para infecção e vias relacionadas com sistema imune eram evidentes no grupo infectado CESC, enquanto nos grupos infectados de LIHC e HNSC o enriquecimento era principalmente relacionado com replicação e reparação de DNA. Este facto parece indicar que a infecção é especialmente ativa no CESC, contradizendo alegações anteriores de que a tumorigenese no colo do útero não estava diretamente ligada à infecção. Nos três tipos de cancro, os vírus integraram os seus genomas no genoma do hospedeiro, afetando a replicação, manutenção e reparação do DNA. No nosso estudo sobre a integração do genoma de HPV16 numa amostra de tumor HNSC, foi confirmada a integração viral no gene humano RAD51B que codifica uma proteína implicada na reparação de DNA por recombinação homóloga. Desta forma, conseguimos confirmar que HPV16 pode atuar tanto como agente cancerígeno directo e indirecto. Provavelmente através da integração do genoma viral no genoma do hospedeiro, a infecção aumentou a quantidade de mutações somáticas no grupo de amostras infectadas em LIHC, mas não em HNSC onde o consumo de tabaco é também um importante agente cancerígeno. O reduzido número de amostras não-infectadas em CESC não permitiu uma comparação fiável da quantidade de mutações somáticas entre grupos de infectados e não-infectados. Ainda assim, nos grupos infectados de LIHC e HNSC, algumas mutações somáticas ocorreram no contexto de vias relacionadas com o sistema imunológico, mostrando que podem contribuir para tornar estes indivíduos susceptíveis à infecção. Além disso, ao verificar a expressão dos genes de HPV16 em cinco amostras de CESC e de HNSC, confirmou-se que os genes E6 e E7 estão entre os mais expressos em muitas das amostras, enquanto que o E2 não é expresso. Os genes E6 e E7 são conhecidos por serem preferencialmente integrados no genoma do hospedeiro, ao contrário do gene E2, o qual controla a expressão daqueles, que não é integrado ou é fragmentado. Acredita-se que é a sobreexpressão de E6 e E7 que inicia a carcinogénese. As taxas de infecção viral inferidas neste trabalho por mining de bases de dados omicos são muito semelhantes aos obtidos pelos métodos tradicionais (Tang et al. 2013), mostrando que a informação disponível nos consórcios internacionais públicos pode elucidar, indiretamente, sobre o envolvimento da infecção viral na tumorigénese. O elevado número de amostras por tumor, a grande variedade de origem geográfica das amostras e a caracterização de alto rendimento para diferentes plataformas omicas permitem comparações e avaliações múltiplas, numa escala não acessível anteriormente.

碩士
國立中興大學
生物科技學研究所
92
Cucumber mosaic virus (CMV) is the type species of the genus Cucumovirus. Some strains of CMV harbor satellite RNAs (satRNAs) that are dependent on CMV for their replication, encapsidation, and transmission. The specific aim of this study is to investigate the 5’- and 3’-terminal signals on satRNAs required for high efficiency replication and to distinguish the effects of nucleotide sequences and structures. The 5’- and/or 3’-terminal 30 nucleotides were replaced by random sequences with synthetic primers in polymerase chain reactions (PCR). The transcripts of mutants were used to inoculate the plants with the helper virus, CMV-NT9. The viabilities of the mutants were analyzed by double-stranded RNA (dsRNA) analyses. The nucleotide sequences of the mutant satRNAs progenies were investigated using RT-PCR followed by DNA sequencing. To differentiate between inoculated transcripts and wild-type satRNA contaminations, a selection marker was introduced into the satRNAs by mutating the Thymine164 into an Adenine to create an NdeI restriction site. Since the 5’- and/or 3’-terminal sequences were randomized, poly(A) polmerase was used to add poly(A) tail at 3’ends of dsRNAs, and oligo(dT) primers were used to amplify the progenies. In addition, the multimeric replication forms of satRNAs were also used as templates for RT-PCR with inverse internal primer pair to amplify the terminal sequences. A viable mutant was recovered from the progenies of 324 inoculation tests. Nucleotide sequencing revealed that the mutant contains an addition of T in the 5’ terminal T stretch region and a deletion of C in the 3’ terminal triple region. The sequence analysis of the multimeric forms revealed that some nucleotides at the junctions between monomers were deleted, and the deletion sites contain similar sequence. The formation of multimers may come from homologous template-switching. Sequence analyses and structural predictions revealed evident differences between viable and non-viable satRNAs. Together, the results indicated that the current wild-type terminal sequences of satRNAs are the most competitive patterns under our experimental conditions. Both structure and sequence may play important roles in the replication of satRNAs.

博士
國防醫學院
生命科學研究所
97
Plants take advantage of the vascular system to operate environmental stimulates for fine-tuning their developmental programs. Recent evidence shows that the FLOWERING LOCUS T (FT) protein is the long-sought-after florigen that integrates the photoperiod variation perceived in the leaves. However, evidence also supports that other yet-to-be identified systemic regulators participate in floral induction. To this end, we investigated phloem exudates from excised broccoli (Brassica oleracea) inflorescences. Microarray and RT-PCR analyses revealed that at least two RNAs of floral regulators, FVE and AGAMOUS-LIKE 24 (AGL24), are present in the phloem sap. Enzymatic analysis demonstrated that the phloem-sap RNAs contain a 5' cap and a polyadenlylation tail, which suggests that phloem sap contains typical mRNAs. Arabidopsis grafting experiments were used to test whether these RNAs move long distance along the phloem translocation stream. Consistent with previous reports, Arabidopsis transformants expressing FVE and AGL24 displayed an early flowering phenotype. When wild-type scions were grafted onto P35S-FVE or P35S-AGL24 transformant stocks, the RNAs of transgenic FVE and AGL24 were detected from the wild-type scions. Thus, both FVE and AGL24 RNAs can move long distance across the graft union. Our data support the notion that multiple systemic floral regulators may participate in floral regulation.

博士
國立中興大學
生物科技學研究所
97
Non-coding RNAs (ncRNAs) play vital roles in translation, splicing, RNA processing, RNA modification and regulation of gene expression. The advancement in ncRNA discovery is evolving along with the finding of new classes of ncRNAs and the invention of revolutionary sequencing platforms. High-throughput sequencing technologies greatly facilitate the study of small regulatory RNAs which are 20 to 30 nt in length. High-throughput sequencing data of 18-26 nt small RNA fragments are a mixture of small regulatory RNAs and degraded products from coding RNAs or ncRNAs. The proper choice of computational approaches in analyzing small RNA sequencing data is crucial for the dissection of small RNAs derived from distinct origins, for making discovery of new ncRNAs and for revealing embedded knowledge in these ncRNAs. To date, the development of computational approaches mostly focused on the discovery of microRNAs (miRNAs). Computational approaches which use small RNA sequencing data for the studies of other ncRNAs are much in need. This dissertation presents the development of novel bioinformatics approaches to analyze small RNA sequencing data and showed that the analyses have increased the understandings of Arabidopsis ncRNAs. In first part, by the use of abundant small RNA sequencing data from the public domain, a new bioinformatics approach was developed for the finding of trans-acting small interfering RNAs (ta-siRNAs), a new class of small regulatory RNAs. Different from that of other siRNAs, the biogenesis of ta-siRNAs is dependent on the cleavage directed by miRNAs. Moreover, most ta-siRNAs are clustered in 21-nt increments relative to the cleavage site. Based on this characteristic, this study developed the first computational algorithm which successfully recovered both known and novel Arabidopsis loci producing ta-siRNAs from complex small RNA sequencing data. A group of newly identified ta-siRNAs was produced by the cleavage directed by a ta-siRNA instead of by miRNAs as was reported previously. The results indicate the existence of a small RNA regulatory cascade initiated by miRNA-directed cleavage and followed by the consecutive production of ta-siRNAs. The second part focuses on the use of small RNA sequencing data in the annotation of small nucleolar RNAs (snoRNAs). Small RNAs from snoRNAs are often considered to be degraded products of snoRNAs and were filtered out without further analysis in previous studies. However, the analysis of Arabidopsis small RNA sequencing data revealed an enrichment of small RNAs at the termini of snoRNAs. With the use of this feature, this study developed a new method which was able to re-annotate known snoRNAs lacking well defined termini and to discover novel snoRNA species. The finding of new snoRNAs also supported that there are additional RNA modification sites on Arabidopsis ribosomal RNAs and spliceosomal small nuclear RNAs. This research demonstrates that, by combining pre-existing biological knowledge and appropriate mining approaches, small RNA sequencing data represent a wealth treasure for the studies of small regulatory RNAs as well as other ncRNAs.

Small RNAs (sRNAs), circa 21-26nt RNA molecules, are a novel class of regulatory molecules that influence many aspects of plant biology. The first objective of this thesis was to utilize computational approaches both to investigate how microRNAs (miRNAs), a type of sRNA, as a class affect their target transcripts’ accumulation and to identify novel miRNAs in Arabidopsis thaliana. The second objective of this thesis was to examine the regulation of protein coding (PC) cis natural antisense transcripts (cis-NATs), which have the potential to make double stranded RNA. Computational analysis of the expression of miRNA-regulated genes demonstrated that the transcriptomes of the inflorescences of plants defective in miRNA biogenesis were similar to normal leaf tissues and dissimilar to normal pollen and seed. Thus, miRNAs cause the plant transcriptome to shift from a vegetative to reproductive state. Known miRNA targets fail to explain miRNA-defective mutant transcriptome patterns. Novel computational approaches were used to discover five new mature miRNAs. Interestingly, two miRNAs have different functions but are encoded by perfect complements of the same precursor molecule. Genome-wide analysis of cis-NAT abundances revealed that protein coding (PC) cis-NATs tend to be co-expressed, broadly expressed, and highly expressed across diverse abiotic stress conditions. These expression patterns were negatively associated with sRNAs because sRNAs were under-represented within PC cis-NATs compared to PC non-cis-NATs. sRNAs also mapped to cis-NATs and non-cis-NATs at similar frequencies in mutants defective in nat-siRNA biogenesis relative to other genotypes. We suggest a common euchromatin environment and possibly antisense RNA stabilization of mRNA transcripts may contribute to the high level, breadth, and co-expression of cis-NATs. However, cis-NATs are correlated less frequently than expected, and cis-NAT transcript abundances often differ more than expected. In addition, sRNAs matched PC cis-NATs relative to PC non-cis-NATs more frequently in abiotic stress conditions than in control conditions. Thus, although sRNAs do not have a widespread role in regulating cis-NATs, sRNAs may have a focused role in regulating cis-NAT transcript abundances.
PhD thesis
NSERC