Les organismes bactériens sont en contact direct avec leur environnement et doivent donc constamment s’acclimater aux variations de celui-ci. Pour cela, plusieurs leviers de régulations peuvent être actionnés. Récemment, la régulation post-transcriptionnelle par les ARN régulateurs a été proposée comme un mécanisme de régulation rapide et peu coûteux pour la cellule. Chez le phytopathogène Dickeya dadantii, la régulation de la virulence a quasi exclusivement été étudiée au niveau transcriptionnel et l’implication des ARN régulateurs dans la virulence reste très peu connue. Pour cela, nous avons tout d’abord étudié le rôle des chaperons à ARN dans la pathogénie de D. dadantii et mis en évidence leur implication dans de nombreux facteurs de virulence comme la production d’enzyme de dégradation de la paroi végétale. Puis, nous avons développé une nouvelle méthode d’identification d’ARN à partir de données RNA-seq. Cette méthode a été développée pour tirer profit des séquençages réalisés en paired-end, permettant de séquencer les deux extrémités d’un transcrit. Son évaluation dans sa capacité à détecter de manière précise des ARN connus a montré une performance supérieure aux méthodes de détection existantes. Enfin, cette nouvelle méthode a été appliquée sur des données de séquençage de petits transcrits. Cette analyse nous a permis d’identifier plus d’un millier d’ARN régulateurs potentiels, dont plusieurs pourraient être impliqués dans la régulation de la virulence. Ces travaux ont donc permis de mettre en lumière l’existence d’une régulation post-transcriptionnelle chez D. dadantii et de proposer des pistes concernant les acteurs et mécanismes concernés / Bacterial organisms are directly exposed to environmental conditions and have to respond to environmental stress. To do so, several regulation network are known. Recently, post transcriptional regulation with small RNAs was suggested to be a fast and cheap in energy regulation mechanism. In the phytopathogen Dickeya dadantii, investigations on pathogenic process mostly focused on its control by transcriptional regulators. Knowledge of post-transcriptional regulation of the virulence factors is still in its infancy.To this end, we first studied the impact of RNA chaperones in the virulence of D. dadantii and showed that they were involved in the regulation of several virulence factors, like production of cell wall degrading enzyme. Then, we developed a new method to detect sRNAs from paired-end bacterial RNA-seq data. This method take paired end sequencing into account, which allow the sequencing of the both ends of each fragment. A comparative assessment showed that this method outperforms all the existing methods in terms of sRNA detection and boundary precision. Finally, this method was applied to sequencing data. With this analysis, more than one thousand sRNAs has been detected, with the identification of several candidates potentially involved in virulence.Thereby, this work highlight the existence of post-transcriptionnal regulation in D. dadantii and suggest candidates and mechanisms involved in this regulation

Many protein structure prediction programs exist and they can efficiently generate a number of protein models of a varying quality. One of the problems is that it is difficult to know which model is the best one for a given target sequence. Selecting the best model is one of the major tasks of Model Quality Assessment Programs (MQAPs). These programs are able to predict model accuracy before the native structure is determined. The accuracy estimation can be divided into two parts: global (the whole model accuracy) and local (the accuracy of each residue). ProQ2 is one of the most successful MQAPs for prediction of both local and global model accuracy and is based on a Machine Learning approach. In this thesis, I present my own contribution to Model Quality Assessment (MQA) and the newest developments of ProQ program series. Firstly, I describe a new ProQ2 implementation in the protein modelling software package Rosetta. This new implementation allows use of ProQ2 as a scoring function for conformational sampling inside Rosetta, which was not possible before. Moreover, I present two new methods, ProQ3 and ProQ3D that both outperform their predecessor. ProQ3 introduces new training features that are calculated from Rosetta energy functions and ProQ3D introduces a new machine learning approach based on deep learning. ProQ3 program participated in the 12th Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction (CASP12) and was one of the best methods in the MQA category. Finally, an important issue in model quality assessment is how to select a target function that the predictor is trying to learn. In the fourth manuscript, I show that MQA results can be improved by selecting a contact-based target function instead of more conventional superposition based functions. / <p>At the time of the doctoral defense, the following paper was unpublished and had a status as follows: Paper 3: Manuscript.</p>