This research investigates the application of machine learning techniques in computational genomics across two distinct domains: (1) the predicting the source of bacterial pathogen using whole genome sequencing data, and (2) the functional annotation of genes using single- cell RNA sequencing data. This work proposes the development of a bioinformatics pipeline tailored for identifying genomic variants, including gene presence/absence and single nu- cleotide polymorphism. This methodology is applied to specific strains such as Salmonella enterica serovar Typhimurium and the Ralstonia solanacearum species complex. Phylo- genetic analyses along with pan-genome and positive selection studiesshow that genomic variants and evolutionary patterns of S. Typhimurium vary across sources, which suggests that sources can be accurately attributed based on genomic variants empowered by machine learning. We benchmarked seven traditional machine learning algorithms, achieving a no- table accuracy of 94.6% in host prediction for S. Typhimurium using the Random Forest model, underscored by SHAP value analyses which elucidated key predictive features. Next, the focus is shifted to the prediction of Gene Ontology terms for Arabidopsis genes using single-cell RNA-seq data. This analysis offers a detailed comparison of gene expression in root versus shoot tissues, juxtaposed with insights from bulk RNA-seq data. The integration of regulatory network data from DAP-seq significantly enhances the prediction accuracy of gene functions. / Master of Science / This work applies machine learning techniques to two areas in computational biology: pre- dicting the hosts of bacterial pathogens based on their genome data, and predicting the func- tions of plant genes using single-cell gene expression data. The first part develops a method to analyze genome sequences from bacterial pathogens like Salmonella enterica serovar Ty- phimurium and the Ralstonia solanacearum species complex, identifying genomic variants, including gene presence/absence and single nucleotide polymorphism, which are variations in genetic code. By studying the evolutionary relationships and genetic diversity among dif- ferent strains, the motivation for using machine learning models to predict the sources (e.g., poultry, swine) of the pathogen genomes is established. Several machine learning models are then trained on these datasets, and the most important factors contributing to the predic- tions are identified. The second part focuses on predicting the functions of genes in the model plant species Arabidopsis thaliana using the gene expression data measured at the single-cell level to train machine learning models for identifying standardized gene function descrip- tions called Gene Ontology (GO) terms. By comparing results from single-cell and bulk tissue data, the study evaluates whether the higher resolution of single-cell data improves gene function prediction accuracy. Additionally, by incorporating information about gene regulation from a specialized experiment, the role of gene expression control in determining gene functions is explored.

Seeds from the desert shrub Simmondsia chinensis (jojoba) are one of the only known natural plant sources to store a majority of its oil in the form of liquid wax esters (WE) instead of triacylglycerols (TAGs) and these oils account for ~55% of the seed weight. Jojoba oil is highly valued as cosmetic additives and mechanical lubricants, yet despite its value much is still unknown about its neutral lipid biosynthetic pathways and lipid droplet packaging machinery. Here, we have used a multi-"omics" approach to study how spatial differences in lipid metabolites, gene expression, and lipid droplet proteins influence the synthesis and storage of jojoba lipids. Through these studies mass spectrometry analyses revealed that WEs are compartmentalized primarily in the cotyledonary tissues, whereas TAGs are, surprisingly, localized to the embryonic axis tissues. To study the differences in gene expression between these two tissues, a de novo transcriptome was assembled from high throughput RNAseq data. Differential gene expression analysis revealed that the Jojoba Wax Synthase, which catalyzes the formation of wax esters, and the Diacylglycerol O-Acyltransferase1, which catalyzes the final acylation of triacylglycerol synthesis, were differentially expressed in the cotyledons and embryonic axis tissues, respectively. Furthermore, through proteomic analysis of lipid droplet proteins from lipid droplets of the cotyledons and embryonic axis, it was estimated that each of these tissues contains a different proportion of the major lipid droplet proteins, oleosins, steroleosins, caleosins, and lipid droplet associated proteins. The Jojoba Olesosin1, Lipid Droplet Associated Protein 1, and Lipid Droplet Associated Protein 3, were identified as potential lipid droplet proteins that could be important for storage of wax esters. The coding sequences of these genes were transiently expressed in N. benthamiana leaves individually, and with co-expression of Mus musculus diacylglycerol acyltransferase 2, and in all cases were able to induce neutral lipid accumulation. These data also suggest a Lipid Droplet Associated Protein 1 has a specialized role for wax ester accumulation in the cotyledons, whereas Lipid Droplet Associated Protein 3 may have a more generalized role for the storage of triacylglycerols. These differences in compartmentation suggests that the cotyledons and embryonic axis of jojoba have evolved tissue-specific sets of genes for neutral lipid assembly and lipid droplet accumulation. It may be important to consider this tissue context for genetic engineering strategies designed to introduce genes from jojoba into other oilseed crops.

A eficiência alimentar (EA) é um fenótipo complexo, controlado por diversos processos biológicos. Determinar e entender esses processos é fundamental para selecionar animais superiores ou mesmo orientar decisões de manejo com o objetivo de aumentar a produtividade e diminuir o impacto ambiental da pecuária. Neste trabalho, propusemos analisar a EA através de uma abordagem de biologia de sistemas, baseada em transcriptômica multitecidual, a fim de gerar um entendimento sistêmico dessa característica. Para isso, 18 animais extremos para consumo alimentar residual foram selecionados a partir de um grupo de 98 bovinos Nelore machos inteiros e tiveram seu transcriptoma de hipotálamo, pituitária, adrenal, músculo e fígado sequenciado (RNAseq). Os reads gerados foram alinhados com o genoma de referência bovino (UMD3.1), filtrados e a expressão de cada gene foi estimada. A partir desses dados três abordagens de análises de dados foram desenvolvidas. Na primeira, cinco critérios de inclusão foram definidos para selecionar genes e construir uma rede de co-expressão para os cinco tecidos, de forma que além de indicarmos diversos genes e processos associados à EA, também fomos capazes de determinar dois genes reguladores, o NR2F6 e o TGFB1. Na segunda abordagem focamos no eixo hipotálamo-pituitária-adrenal, também utilizando análises de co-expressão, mas dessa vez sem partir de prévia seleção de genes e concluímos que o sistema de recompensa do cérebro pode estar envolvido no estímulo para maior consumo de alimentos observado no grupo de baixa EA. Finalmente, com a terceira abordagem, identificamos RNAs longos não codificadores (lncRNAs) expressos nos cinco tecidos e encontramos 30 transcritos expressos diferencialmente entre a alta e baixa EA na pituitária, músculo e adrenal, sendo que alguns deles se mostraram relacionados a processos já previamente demostrados como sendo associados a essa característica. Concluímos que, apesar de não conseguirmos determinar nesse momento o papel da maior susceptibilidade ao estresse, reportado na literatura para animais de baixa EA, no estímulo para maior ingestão de alimentos desse grupo, o sistema de recompensa hipotalâmico parece estar envolvido nesse processo. A maior ingestão pode ser a causa da resposta inflamatória observada no fígado, sendo ela de origem bacteriana, indicada pela maior concentração de endotoxina sérica nos animais menos eficientes. O maior turnover de proteínas no músculo de animais de baixa EA já havia sido indicado como um dos fatores que levam ao maior gasto energético nesses indivíduos e foi confirmado nesse trabalho. Além de alguns fatores de transcrição serem indicados como reguladores centrais desse fenótipo, lncRNAs também parecem ter função regulatória importante na EA. / Feed efficiency (FE) is a complex phenotype, controlled by several biological processes. Determining and understanding these processes is fundamental to select superior animals or even guide management decisions, aiming to increase productivity and reduce the environmental impact of livestock. In this work, we propose to analyze FE through a systems biology approach, based on multi-tissue transcriptomics, in order to generate a systemic understanding of this trait. For this purpose, 18 extreme animals for residual feed intake were selected from a group of 98 male Nellore cattle and had their hypothalamus, pituitary, adrenal gland, muscle and liver transcriptome sequenced (RNAseq). Reads generated were aligned with the bovine reference genome (UMD3.1), filtered and the expression of each gene was estimated. From these data three experiments were developed. In the first one, five inclusion criteria were defined to select genes and to construct a network of coexpression for the five tissues, so that besides indicating several genes and processes associated with EA, we were also able to determine two regulatory genes, NR2F6 and TGFB1. In the second experiment, we focused on the hypothalamic-pituitary-adrenal axis, also using co-expression analysis, but this time without starting from previous selected genes. We conclude that the reward system of the brain might be involved in the stimulus for higher feed intake observed in the low EA group. Finally, in the third experiment, we identified long noncoding RNAs (lncRNAs) expressed in the five tissues and found 30 transcripts differentially expressed between the high and low FE in the pituitary, muscle and adrenal, and some of them were related to previously demonstrated processes associated to this trait. We conclude that although we cannot determine at this time the role of higher susceptibility to stress, reported in the literature for animals of low FE, in the stimulus for higher feed intake of this group, the hypothalamic reward system seems to be involved in this process. The higher ingestion might be the cause of the inflammatory response observed in the liver of, being of bacterial origin, indicated by the higher concentration of serum endotoxin in less efficient animals. The higher turnover of proteins in the muscle of low FE animals had already been indicated as one of the factors that lead to higher energy expenditure in these individuals and it was confirmed in this study. In addition to some transcription factors being indicated as central regulators of this phenotype, lncRNAs also appear to play an important regulatory role in FE.

La méthylation de l'ADN est une marque épigénétique répressive impliquée dans de nombreux processus physiologiques et pathologiques. Des hyperméthylations de promoteurs sont ainsi responsables de répressions transcriptionnelles de gènes suppresseurs de tumeurs dans les cancers. La méthylation de l'ADN serait capable d'induire une répression transcriptionnelle par la combinaison de deux mécanismes principaux : l'éloignement de facteurs de transcription activateurs, et le recrutement de protéines répressives liant spécifiquement l'ADN méthylé. MBD2 est une protéine de liaison à l'ADN méthylé capable de recruter les complexes répresseurs NuRD et SIN3A. ZBTB4 est capable de se lier à l'ADN méthylé in vitro et induit une répression de la transcription de plasmides méthylés lorsqu'elle est surexprimée. Son rôle de répresseur transcriptionnel dépendant de la méthylation de l'ADN reste toutefois peu documenté. Nous avons identifiés par RNAseq les modifications du transcriptome induites par une déplétion de MBD2 ou de ZBTB4. Les gènes surexprimés après déplétion de MBD2 ou ZBTB4 sont méthylés sur leur promoteur, et sont aussi surexprimés après traitement avec des agents déméthylants. Des résultats d'immuno-précipitations de chromatine réalisées contre les deux protéines endogènes montrent que la quasi-totalité des sites de fixation de MBD2 et qu'une partie des sites de fixations de ZBTB4 correspondent à des régions méthylés. Ces résultats confirment à l'échelle du génome que MBD2 endogène est bien un interprète majeur de la méthylation de l'ADN, et que ZBTB4 réprime bien la transcription de gènes méthylés. Nous avons aussi observé une redistribution importante de MBD2 sur le génome dans des modèles de progression tumorale. Nos résultats montrent que les gènes réprimés pendant la transformation oncogénique le sont en partie par MBD2. L'expression de certains de ces gènes peut être induite dans les lignées transformées par déplétion de MBD2 par siRNA / DNA methylation is an epigenetic mark that plays a role in many physiological and pathological processes. Indeed, silencing of tumor suppressor genes in cancer is frequently caused by promoter hypermethylations. Transcriptional repression induced by DNA methylation is likely caused by the combination of two mechanisms: the repulsion of activator transcription factors, and the recruitment of repressor proteins able to specifically recognize methylated DNA. MBD2 is a methyl DNA binding protein that cans recruits NuRD or SIN3A repressor complexes. ZBTB4 is able to bind methylated DNA in vitro, and can repress the transcription of methylated plasmids when overexpressed. Its methylationdependent transcriptional repressor function remains poorly documented. By RNAseq, we have identified transcriptomic modifications induced by the depletion of either MBD2 or ZBTB4. Genes up regulated after MBD2 or ZBTB4 depletion were methylated on their promoter, and were also up regulated after treatment with demethylating agents. Chromatin immunoprecipitations experiments against endogenous proteins showed that almost all MBD2 binding sites, and that a part of ZBTB4 binding sites, correspond to methylated DNA regions. These results confirmed at genome wide scale that endogenous MBD2 is a major reader of DNA methylation and that ZBTB4 does repress the transcription of methylated genes. We observed an important redistribution of MBD2 on the genome in models of tumor progression. Our results showed that MBD2 plays role in gene repressions occurring during oncogenic transformation. Some of those repressed genes can be re-expressed in transformed cell lines after depletion of MBD2 by siRNA

Beef quality in Brazil is important for both consumers and the food industry due to high demand and competitiveness in the domestic and international markets. Therefore, it is necessary to develop research to improve beef quality of Nellore cattle (Bos indicus), mainly tenderness, one of the main features to add value to meat. New-generation technologies provide accurate, rapid and inexpensive information on the entire genome, showing great advantage over conventional methods for sequencing and gene expression. However, these new technologies generate large database, which require the use of bioinformatics tools for data analyses of sequencing and for a better understanding of biological regulation mechanisms , cellular control, gene interactions, among other applications. In a previous study, samples were collected from the Longissimus dorsi muscle of 790 animals from Nellore cattle and shear force assessments were made 24 hours after slaughter, with seven and 14 days of aging. Aiming to identify differentially expressed (DE) genes, 34 samples from Nellore animals with extreme levels of estimated breeding value (EBV) for shear force (SF) were selected, sequenced by the method of RNA sequencing (RNA-Seq) (Illumina HiScanSQ). This study performed the processing of data generated by RNA-Seq using software QuasiSeq and Cuffdiff. In the QuasiSeq analysis, 22 DE genes were found, while in the Cuffdiff analysis, 113 DE genes were found. To better understand the biological process involved in meat tenderness, integrative analysis identified possible regulators that can explain the activity of transcriptional regulation in this process using partial correlation coefficient with information theory (PCIT), phenotypic impact factor (PIF) and regulatory impact factor (RIF) methods. The genes found in the PCIT analysis USP2, GBR10, ANO1 and TMBIM4; microRNAs found in RIF analysis bta-mir-133a-2 and bta-mir-22, and the genes with high PIF value MB, ENO3, CA3 could be fundamental to unravel the complex molecular mechanisms that control the meat tenderness in Nellore cattle. / A qualidade da carne bovina no Brasil é importante tanto para o consumidor, como para a indústria alimentícia devido à alta competitividade e exigência do mercado nacional e internacional. Portanto, é necessário o desenvolvimento de pesquisas para melhorar a qualidade da carne bovina da raça Nelore (Bos indicus), principalmente a maciez, que é considerada uma das principais responsáveis por agregar valor à carne. Tecnologias de nova geração proporcionam informações precisas, rápidas e baratas de todo genoma, mostrando grande vantagem em relação aos métodos convencionais de sequenciamento e de estudos de expressão gênica. Essas novas tecnologias geram um grande volume de dados, sendo necessário o uso de ferramentas de bioinformática para realizar as análises de sequenciamento e ter uma maior compreensão de mecanismos biológicos de regulação, controle celular, interações gênicas, entre outras aplicações. Em um estudo prévio, foram coletadas amostras do músculo Longissimus dorsi de 790 animais da raça Nelore e foram realizadas avaliações da força de cisalhamento 24 horas após abate, e com sete e 14 dias de maturação. Com o objetivo de identificar genes diferencialmente expressos (DE), foram selecionadas no total 34 amostras de animais da raça Nelore com valores extremos de valor genético estimado (EBV) para força de cisalhamento (SF), e sequenciados pelo método de sequenciamento de RNA (RNA-Seq) (Illumina HiScanSQ). Neste estudo foi realizado o processamento dos dados gerados pelo RNA-Seq através dos softwares QuasiSeq e Cuffdiff. Foram encontrados 22 genes DE para as análises do QuasiSeq e 113 genes DE para as análises do Cuffdiff. Para melhor compreensão dos processos biológicos envolvidos na maciez da carne, análises integrativas identificaram possíveis reguladores que podem explicar a atividade de regulação transcricional neste processo utilizando os métodos do Coeficiente de Correlação Parcial com Teoria da Informação (PCIT), Fator de Impacto Fenotípico (PIF) e Fator de Impacto Regulatório (RIF). Os genes encontrados nas análises análises do PCIT USP2, GBR10, ANO1 e TMBIM4, assim como os microRNAs encontrados nas análises do RIF, bta-mir-133a-2 e bta-mir-22 e os genes de maior valor de PIF MB, ENO3, CA3 podem ser fundamentais para desvendar os complexos mecanismos moleculares que controlam a maciez da carne na raça Nelore.

Gene expression is tightly regulated by complex transcriptional regulatory mechanisms to achieve specific expression patterns, which are essential to facilitate important biological processes such as embryonic development. Dysregulation of gene expression can lead to diseases such as cancers. A better understanding of the transcriptional regulation will therefore not only advance the understanding of fundamental biological processes, but also provide mechanistic insights into diseases. The earlier versions of high-throughput expression profiling techniques were limited to measuring average gene expression across large pools of cells. In contrast, recent technological improvements have made it possible to perform expression profiling in single cells. Single-cell expression profiling is able to capture heterogeneity among single cells, which is not possible in conventional bulk expression profiling. In my PhD, I focus on developing new algorithms, as well as benchmarking and utilising existing algorithms to study the transcriptomes of various biological systems using single-cell expression data. I have developed two different single-cell specific network inference algorithms, BTR and SPVAR, which are based on two different formalisms, Boolean and autoregression frameworks respectively. BTR was shown to be useful for improving existing Boolean models with single-cell expression data, while SPVAR was shown to be a conservative predictor of gene interactions using pseudotime-ordered single-cell expression data. In addition, I have obtained novel biological insights by analysing single-cell RNAseq data from the epiblast stem cells reprogramming and the leukaemia systems. Three different driver genes, namely Esrrb, Klf2 and GY118F, were shown to drive reprogramming of epiblast stem cells via different reprogramming routes. As for the leukaemia system, FLT3-ITD and IDH1-R132H mutations were shown to interact with each other and potentially predispose some cells for developing acute myeloid leukaemia.

No description available.

Submitted by Maria Beatriz Vieira (mbeatriz.vieira@gmail.com) on 2016-08-03T13:55:47Z No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) tese_pablo_cadore.pdf: 5773635 bytes, checksum: 63fadc0a8749af862e0e0b7a9b075c0c (MD5) / Made available in DSpace on 2016-08-03T13:55:47Z (GMT). No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) tese_pablo_cadore.pdf: 5773635 bytes, checksum: 63fadc0a8749af862e0e0b7a9b075c0c (MD5) Previous issue date: 2015-09-28 / Sem bolsa / O arroz é uma cultura de origem tropical amplamente cultivada em uma diversidade de áreas, atualmente é o segundo cereal mais produzido no mundo. A ocorrência de baixas temperaturas é um estresse comum na cultura do arroz em regiões temperadas, portanto a tolerância a baixas temperaturas é uma característica desejável em genótipos brasileiros de arroz cultivados no sul do país, onde as temperaturas baixas prejudicam a germinação, o estabelecimento da lavoura e diminuem o rendimento de grãos. Portanto, este trabalho tem como objetivo estudar o comportamento de cruzamentos de arroz tolerantes e sensíveis para o caráter de tolerância ao frio no estádio de germinação e identificar genes diferencialmente expressos através do transcriptoma. Foram realizados cruzamentos controlados com 5 genitores contrastantes. Os genitores e as populações segregantes em F3 e F4 foram submetidos a tratamentos com diferentes temperaturas (13C e 25C) e comparados quanto ao seu desempenho relativo, medido pela germinação, índice de velocidade de germinação em laboratório e em campo, comprimento do coleóptilo e comprimento da raiz. A análise do transcriptoma foi realizada através do RNAseq de dois genótipos BRS SCS Tio Taka e Oro, submetidos a tratamentos com diferentes temperaturas (13C e 25C). Os resultados indicam que o índice de velocidade de germinação obtido em laboratório, apresenta maior eficiência em identificar genótipos e populações comparado com a avaliação em campo. O comprimento do coleóptilo e o índice de velocidade de germinação são as metodologias mais adequadas pela elevada associação com a germinação em baixas temperaturas. Os resultados das correlações na geração F3 diferem dos resultados apresentados na geração F4, devido a segregação das populações. Na cultivar tolerante Oro, poucos genes tiveram expressão gênica aumentada significativamente. Uma lista de 55 genes com diferença na expressão comuns as duas cultivares e uma lista com 17 genes com diferença na expressão únicos para a cultivar tolerante Oro, foram selecionados devido a respostas desses genes à tolerância ao frio na germinação de arroz. / Rice is a tropical crop widely grown in a variety of areas, is currently the second most-produced cereal in the world. The presence of low temperatures is a common stress in rice crops in temperate regions, so the tolerance to low temperatures is a desirable characteristic in Brazilian rice genotypes cultivated in the south, where low temperatures damage germination, establishment of the crop and decrease grain yield. Therefore, this paper aims to study the behavior of rice crossings tolerant and sensitive to cold tolerance at the germination stage and identify differentially expressed genes through the transcriptome. Controlled crosses with 5 contrasting genotypes were performed. The parents and segregating populations in F3 and F4 were subjected to treatments with different temperatures (13C e 25C) and compared for their relative performance, as measured by germination, germination speed index in the laboratory and in the field, length coleoptile length and root. The transcriptome analysis was performed through the RNAseq two genotypes BRS SCS Tio Taka and Oro, submitted to treatment with different temperatures (13C e 25C). The results show that the germination speed index obtained in the laboratory, has higher efficiency and populations to identify genotypes compared with the field evaluation. The length of the coleoptile and the germination speed index are the most appropriate methodologies due the high association with germination at low temperatures. The results of the correlations in the generation F3 differed from the results shown in the F4 generation due to segregation populations. In tolerant cultivar Oro, few genes have increased gene expression significantly. A list of 55 genes with common expression difference in the two cultivars and a list of 17 genes with only difference in the expression for the tolerant cultivar Oro, were selected due to responses of these genes to cold tolerance in rice germination.

Beef quality in Brazil is important for both consumers and the food industry due to high demand and competitiveness in the domestic and international markets. Therefore, it is necessary to develop research to improve beef quality of Nellore cattle (Bos indicus), mainly tenderness, one of the main features to add value to meat. New-generation technologies provide accurate, rapid and inexpensive information on the entire genome, showing great advantage over conventional methods for sequencing and gene expression. However, these new technologies generate large database, which require the use of bioinformatics tools for data analyses of sequencing and for a better understanding of biological regulation mechanisms , cellular control, gene interactions, among other applications. In a previous study, samples were collected from the Longissimus dorsi muscle of 790 animals from Nellore cattle and shear force assessments were made 24 hours after slaughter, with seven and 14 days of aging. Aiming to identify differentially expressed (DE) genes, 34 samples from Nellore animals with extreme levels of estimated breeding value (EBV) for shear force (SF) were selected, sequenced by the method of RNA sequencing (RNA-Seq) (Illumina HiScanSQ). This study performed the processing of data generated by RNA-Seq using software QuasiSeq and Cuffdiff. In the QuasiSeq analysis, 22 DE genes were found, while in the Cuffdiff analysis, 113 DE genes were found. To better understand the biological process involved in meat tenderness, integrative analysis identified possible regulators that can explain the activity of transcriptional regulation in this process using partial correlation coefficient with information theory (PCIT), phenotypic impact factor (PIF) and regulatory impact factor (RIF) methods. The genes found in the PCIT analysis USP2, GBR10, ANO1 and TMBIM4; microRNAs found in RIF analysis bta-mir-133a-2 and bta-mir-22, and the genes with high PIF value MB, ENO3, CA3 could be fundamental to unravel the complex molecular mechanisms that control the meat tenderness in Nellore cattle. / A qualidade da carne bovina no Brasil é importante tanto para o consumidor, como para a indústria alimentícia devido à alta competitividade e exigência do mercado nacional e internacional. Portanto, é necessário o desenvolvimento de pesquisas para melhorar a qualidade da carne bovina da raça Nelore (Bos indicus), principalmente a maciez, que é considerada uma das principais responsáveis por agregar valor à carne. Tecnologias de nova geração proporcionam informações precisas, rápidas e baratas de todo genoma, mostrando grande vantagem em relação aos métodos convencionais de sequenciamento e de estudos de expressão gênica. Essas novas tecnologias geram um grande volume de dados, sendo necessário o uso de ferramentas de bioinformática para realizar as análises de sequenciamento e ter uma maior compreensão de mecanismos biológicos de regulação, controle celular, interações gênicas, entre outras aplicações. Em um estudo prévio, foram coletadas amostras do músculo Longissimus dorsi de 790 animais da raça Nelore e foram realizadas avaliações da força de cisalhamento 24 horas após abate, e com sete e 14 dias de maturação. Com o objetivo de identificar genes diferencialmente expressos (DE), foram selecionadas no total 34 amostras de animais da raça Nelore com valores extremos de valor genético estimado (EBV) para força de cisalhamento (SF), e sequenciados pelo método de sequenciamento de RNA (RNA-Seq) (Illumina HiScanSQ). Neste estudo foi realizado o processamento dos dados gerados pelo RNA-Seq através dos softwares QuasiSeq e Cuffdiff. Foram encontrados 22 genes DE para as análises do QuasiSeq e 113 genes DE para as análises do Cuffdiff. Para melhor compreensão dos processos biológicos envolvidos na maciez da carne, análises integrativas identificaram possíveis reguladores que podem explicar a atividade de regulação transcricional neste processo utilizando os métodos do Coeficiente de Correlação Parcial com Teoria da Informação (PCIT), Fator de Impacto Fenotípico (PIF) e Fator de Impacto Regulatório (RIF). Os genes encontrados nas análises análises do PCIT USP2, GBR10, ANO1 e TMBIM4, assim como os microRNAs encontrados nas análises do RIF, bta-mir-133a-2 e bta-mir-22 e os genes de maior valor de PIF MB, ENO3, CA3 podem ser fundamentais para desvendar os complexos mecanismos moleculares que controlam a maciez da carne na raça Nelore.

Il devient de plus en plus clair que pour traiter efficacement les tumeurs solides, nous devons également nous intéresser au microenvironnement tumoral. Physiologiquement, les zones intratumorales peuvent présenter une disponibilité anormale en nutriments, un pH altéré ou encore des niveaux d’oxygène bas (hypoxie). Il est connu que l’adaptation hypoxique engendre des cellules tumorales qui sont plus difficiles à traiter indépendamment de l’approche thérapeutique. De plus, l’adaptation hypoxique est nécessaire pour la progression tumorale puisque cette dernière favorise des processus tels que: la survie cellulaire, la motilité, l’angiogenèse, le métabolisme du glucose, l’immunomodulation ainsi que la résistance aux médicaments. Ces phénotypes passent par la régulation des ARN messager (ARNm) et des micro ARN (miARN). Pour ces raisons, des efforts importants ont été déployés pour comprendre l’adaptation hypoxique et les interventions thérapeutiques potentielles pouvant la contrer. À l’heure actuelle, il y a un manque de cohérence et de variété dans les protocoles de traitement hypoxique in vitro qui ne tient pas compte des aspects importants de l’hypoxie in vivo, comme la réduction de la disponibilité en éléments nutritifs, la durée de l’exposition hypoxique ainsi que le degré d’hypoxie. Pour mieux simuler le microenvironnement hypoxique in vivo, nous avons développé de nouveaux protocoles hypoxiques in vitro qui visent à simuler ces aspects. Tout d’abord, en utilisant une lignée cellulaire B16-HIF1a-eGFP, nous avons optimisé le stress métabolique à court terme en conjonction avec l’hypoxie pour augmenter la stabilisation de l’HIF1a. Pour déterminer comment le programme HIF1 adapte les cellules à ces différentes conditions, nous avons analysé les données de séquencage d’ARN qui démontrent que le stress métabolique induit un programme transcriptionnel HIF1 plus robuste et diversifié dans les cellules hypoxiques, et que ce dernier est représentatif du stress hypoxique in vivo. Nous avons également identifié de nouveaux miARN induits par l’hypoxie et démontré que notre protocole d’incubation régule davantage les miRNA associés au pronostic négatif du patient. Nous avons aussi étudié l’adaptation hypoxique à long terme et extrême in vitro. Nous avons observé que l’incubation hypoxique à long terme induit une transition épithéliale à mésenchymateuse (TME), indépendante de l’expression différentielle des facteurs de transcription du TME canonique. Ce changement se produit à des niveaux spécifiques d’oxygène, et nécessite une pré-incubation à des niveaux hypoxiques plus faible. Avec ce protocole, nous avons découvert une nouvelle isoforme de WT1 (tWT1), un moteur potentiel du TME. tWT1 commence la transcription dans l’intron 5 du gène WT1, une région avec plusieurs séquences d’ADN contenant des éléments de réponse à l’hypoxie. La protéine tWT1 a une fonctionnalité limitée : elle est localisée au niveau du noyau, et conserve la liaison de l’ADN aux régions précédemment connues. Nous avons aussi identifié l’expression de tWT1 dans les échantillons de patients atteints de leucémie ainsi qu’une isoforme tWT1 potentiellement plus fonctionnelle grâce à des analyses par kmer. Pour cibler ces phénotypes identifiés dans nos expériences d’adaptation hypoxiques, nous avons développé une nouvelle catégorie d’ARN intérférent (ARNi) thérapeutique : le microARN synthétique (synmiR). Les synmiR sont des molécules de RNAi avec des multiples cibles. En utilisant des expériences in vivo, nous avons établi de nouveaux principes de RNAi qui élargissent considérablement l’espace de conception pour les synmiR. Nous avons mis au point deux algorithmes de conception de synmiR distincts et avons testé leur efficacité dans le contrôle de l’activité transcriptionnelle du génome du VIH in vivo. En conclusion, nous avons montré que l’inclusion de facteurs physiologiques supplémentaires associés à l’hypoxie in vitro entraîne un engagement plus robuste de l’adaptation de l’hypoxie. À ce jour, aucun de nos protocoles d’hypoxie n’a été reproduit dans la littérature. Nous contribuons aux connaissances dans le domaine en décrivant les nouveaux ARNm/miARN induits par l’hypoxie, ainsi que la méthode d’induction fiable de l’EMT par l’hypoxie seulement. Nous faisons également état de l’existence de nouveaux isoformes de WT1 et de leurs liens avec le cancer et l’hypoxie. La connaissance de ces isoformes est importante pour l’avenir de la recherche sur WT1, car elle pourrait faire la lumière sur des résultats auparavant inexplicables. Notre travail dans les synmiR ouvre une nouvelle voie d’investigation pour le traitement de certaines maladies, et fournit un mécanisme d’action testable pour les miRNA endogènes. Une fois suffisamment développés, les synmiR offrent une occasion thérapeutique unique d’exploiter leur multi-ciblage pour avoir un impact spectaculaire sur une seule voie, ou affecter plusieurs voies par le ciblage simultané de gènes clés. / It is becoming increasingly clear that in order to effectively treat solid tumours, we must also address the tumour microenvironment. Physiologically, intratumoral areas may have abnormal nutrient availability, pH, or lower oxygen levels (hypoxia). It is known that hypoxic adaptation results in tumour cells which are harder to treat regardless of therapeutic approach, and hypoxic adaptation is necessary for disease progression due to the induction of tumour promoting phenotypes such as, but not limited to: cell survival, motility, angiogenesis, glucose metabolism, immunomodulation, and drug resistance. This is accomplished through the regulation of both mRNAs and miRNAs. For these reasons, significant effort has been applied to understanding hypoxic adaptation and potential therapeutic interventions. Currently, there is a lack of consistency and protocol variety in in vitro hypoxic treatments that leaves out important aspects of in vivo hypoxia, such as reduced nutrient availability, length of hypoxic exposure, and degree of hypoxia. To better simulate the in vivo hypoxic microenvironment, we have developed new in vitro hypoxic protocols which aim to simulate these aspects. First, using a B16-HIF1α-eGFP hypoxia reporter cell line, we optimized short-term metabolic stress in conjunction with hypoxia to enhance HIF1α stabilization. To ascertain how the HIF1 program adapts the cells to these different conditions, deep transcriptome profiling were performed and demonstrated metabolic stress induces a more robust and diversified HIF1 transcriptional program in cells under hypoxia, which was more representative of in vivo hypoxic stress. We identified novel hypoxia-induced miRNAs as well, and demonstrated our incubation protocol regulated more miRNAs associated with negative patient prognosis. We also investigated long-term and extreme hypoxic adaptation in vitro. Long term hypoxic incubation induced a epithelial to mesenchymal transition (EMT), independent of canonical EMT factor differential expression. This switch occurred at specific oxygen levels, and required pre-incubation at milder hypoxic levels, highlighting the relevance of simulating in vivo hypoxia development in vitro. Through this protocol, we discovered a novel isoform of WT1 (tWT1), a potential driver of our EMT. tWT1 begins transcription within intron 5 of the WT1 gene, a region with several Hypoxia Response Elements DNA sequences. tWT1 retains limited functionality: it is able to localize to the nucleus, and retains DNA binding to previously known gene promoter regions. We also identified the expression of tWT1 in leukemic patient samples as well as a potentially more functional tWT1 isoform through kmer-based analyses. To target these multiple phenotypes identified in our hypoxia adaptation experiments, we worked towards developing a new category of RNA-interference (RNAi) therapeutic, the synthetic microRNA (synmiR). SynmiRs are single-sequence, multi-targeted RNAi molecules. Using in vivo knock-down experiments, we established new RNAi principles which dramatically expand the design space for synmiRs. We developed two philosophically distinct synmiR design algorithms, and validated their efficacy in controlling HIV genome transcriptional activity in vivo. In conclusion, we have shown the inclusion of additional physiological factors associated with hypoxia in vitro results in a more robust engagement of hypoxia adaptation. To date, neither of our hypoxia protocols have been replicated in the literature. We contribute to the literature by describing novel hypoxia induced mRNAs/miRNAs, as well as methods for reliably inducing EMT through hypoxia alone. We also discovered the existence of novel WT1 isoforms and their links to cancer and hypoxia. Knowledge of these isoforms is important for WT1 research moving forward, as it may shed light on previously unexplainable results. Our work in synmiRs opens a new therapeutic avenue for multiple disease states, and provides a testable mechanism of action for endogenous miRNAs. Once sufficiently developed, synmiRs offer a unique therapeutic opportunity to harness their multi-targeting to dramatically impact a single pathway, or affect multiple pathways through simultaneous targeting of key genes.