Dans le domaine de l'ingénierie du logiciel, les systèmes de types sont souvent considérés pour la prévention de l'occurrence de termes dénués de sens par rapport à une spécification des types. Dans le cadre de l'extension d'un langage de programmation avec des caractéristiques dédiées, le typage de ces dernières doit être compatible avec les caractéristiques du langage hôte. Cette thèse se situe dans le contexte de la réécriture de termes embarquée dans la programmation orientée objet. Elle vise à développer un système de types avec sous-typage pour le support du filtrage de motifs associatif sur des termes algébriques construits sur des opérateurs variadiques. Ce travail s'appuie sur le langage de réécriture Tom qui fournit des constructions de filtrage de motifs et des stratégies de réécriture à des langages généralistes comme Java. Nous décrivons l'évaluation de code Tom à travers la définition de la sémantique opérationnelle de ce langage en tant qu'élément essentiel de la preuve de la sûreté du système de types. Celui-ci inclut la vérification de types ainsi que l'inférence de types à base de contraintes. Le langage de contraintes est composé d'une part, de contraintes d'égalité, résolues par unification, d'autre part, de contraintes de sous-typage, résolues par la combinaison de phases de simplification, de génération d'une solution et de ramassage de miettes. Le système de types a été intégré au langage Tom, ce qui permet une plus forte expressivité et plus de sûreté a fin d'assurer que les transformations décrites par des règles de réécriture préservent le type des termes / In software engineering, type systems are often considered in order to prevent the occurrence of meaningless terms in regard to a type specification. When extending a given programming language with new dedicated features, the typing of these features must be compatible with the ones in the host language. This thesis is situated in the context of term rewriting embedded in object-oriented programming and aims to develop a safe type system featuring subtyping for the support of associative pattern matching on algebraic terms built from variadic operators. In this work we consider the Tom rewriting language that provides associative pattern matching constructs and rewrite strategies for Java. We describe Tom code evaluation through the definition of the operational semantics of the Tom language as an essential element to show that the type system is safe. The type system includes type checking and constraint-based type inference. The constraint language is composed of equality constraints solved by unification and subtyping constraints solved by a combination of simplification, generation of solution and garbage collecting. The type system was integrated in Tom which provides both stronger expressiveness and more safety able to ensure that the transformations described by rewrite rules preserve the type of terms

A epidemia da Síndrome de Imunodeficiência Adquirida (AIDS) fez aumentar a incidência de sarcoma de Kaposi (SK) em todos os países e o SK passou a ser considerado doença definidora de AIDS. Desde a descoberta de seu agente etiológico, o herpesvírus humano tipo 8 (HHV-8), vários estudos vêm sendo realizados com o objetivo de caracterizar subtipos virais presentes em todas as formas de SK: clássica, endêmica, iatrogênica e epidêmica. Os sistemas rotineiramente usados na subtipagem do HHV-8 têm usado o seqüenciamento de um gene que contém regiões hipervariáveis e que codifica uma glicoproteína de membrana viral (ORF K1). O presente trabalho apresenta um sistema de subtipagem alternativo que se baseia na presença de sítios de restrição enzimática em um pequeno segmento do gene ORF K1, região hipervariável 1 (VR1) e que permite discriminar subtipos virais. A análise de 68 seqüências; 50 que pertenciam a 36 pacientes com sarcoma de Kaposi infectados e não infectados pelo HIV-1 de São Paulo e 18 a protótipos dos subtipos A a E, mostraram mapas de restrição enzimática característicos dos principais subtipos virais descritos até o momento. Tomando como base apenas as enzimas disponíveis comercialmente, foram selecionadas cinco que se mostraram úteis para a subtipagem de HHV-8: Taq I, Nsi I, Hinf I, Hae III e Mse I. Os resultados obtidos com a técnica de PCR-RFLP (reação em cadeia de polimerase associada à análise do polimorfismo de fragmentos de restrição enzimática) mostraram que de 48 espécimes brasileiros previamente classificados como sendo dos subtipos A, B e C por seqüenciamento gênico, todos foram corretamente subtipados pela técnica de PCR-RFLP. Três amostras (duas do subtipo A e uma do B) apresentaram mais um sítio de restrição enzimática além dos descritos como sendo os predominantes. Mais recentemente, outras 27 amostras de 18 casos de infecção por HHV-8 foram subtipadas pela PCR-RFLP. Houve detecção de oito isolados do subtipo A, sendo seis de variante predominante, um de variante minoritária conhecida e um de nova variante viral. Dois casos de infecção por HHV-8 do subtipo B e sete do subtipo C também foram identificados. Finalmente, um provável caso de infecção pelo subtipo E foi encontrado em paciente com SK-AIDS disseminado, resistente à quimioterapia. Na avaliação global, houve maior número de casos de infecção por HHV-8 dos subtipos A e C. Concluindo, devido à alta sensibilidade e especificidade, baixo custo, rapidez e facilidade de execução, a técnica de PCR-RFLP pode ser usada em larga escala para estudos de epidemiologia molecular, principalmente em países em desenvolvimento. / AIDS epidemic has increased the incidence rates of Kaposi’ s sarcoma (KS) in all countries, and KS has been considered an AIDS-defining illness. Since the discovery of the human herpesvirus 8 (HHV-8), the etiological agent of KS, several studies have been conducted in order to characterize HHV-8 in all forms of KS: classic, endemic, iatrogenic, and epidemic. The HHV-8 genome presents a hypervariable region termed ORF K1 useful for virus subtyping. The objectives of the present study were to describe an alternative method for subtyping HHV-8, to compare this new method with DNA sequencing, and to use this method for HHV-8 subtyping. After cloning and sequencing a segment of the ORF K1 (VR1) in 50 HHV-8/DNA isolates from 36 Brazilian KS-AIDS patients, we searched for restriction enzymatic sites in this segment of DNA, and compared them with 18 sequences reported in the literature. Then we constructed the enzymatic restriction maps useful for discriminating all HHV-8 subtypes described up to now, and standardized a PCR-RFLP (restriction fragment length polymorphism analysis) using five commercial enzymes: Taq I, Nsi I, Hinf I, Hae III e Mse I. After comparing the results obtained by the two methods, we used PCR-RFLP for HHV-8 subtyping in 27 new HHV-8/DNA isolates. The results obtained by DNA sequencing and PCR-RFLP showed 100% of concordance, and allowed the use of PCR-RFLP for HHV-8 subtyping. Indeed, we disclosed that among KS-AIDS patients from São Paulo, subtypes A and C are more prevalent than subtype B. Although additional restriction sites were detected in some Brazilian HHV-8 isolates, the majority of them belonged to the predominant strains described in the literature. Interestingly, one probable case of HHV-8 subtype E was detected in a patient who presented disseminated KS and resistance to chemotherapy. Because of its high sensitivity, specificity, low cost, and rapid execution, PCR-RFLP could be used on a large scale, mostly in countries with poor resources and where KS is endemic.

A epidemia da Síndrome de Imunodeficiência Adquirida (AIDS) fez aumentar a incidência de sarcoma de Kaposi (SK) em todos os países e o SK passou a ser considerado doença definidora de AIDS. Desde a descoberta de seu agente etiológico, o herpesvírus humano tipo 8 (HHV-8), vários estudos vêm sendo realizados com o objetivo de caracterizar subtipos virais presentes em todas as formas de SK: clássica, endêmica, iatrogênica e epidêmica. Os sistemas rotineiramente usados na subtipagem do HHV-8 têm usado o seqüenciamento de um gene que contém regiões hipervariáveis e que codifica uma glicoproteína de membrana viral (ORF K1). O presente trabalho apresenta um sistema de subtipagem alternativo que se baseia na presença de sítios de restrição enzimática em um pequeno segmento do gene ORF K1, região hipervariável 1 (VR1) e que permite discriminar subtipos virais. A análise de 68 seqüências; 50 que pertenciam a 36 pacientes com sarcoma de Kaposi infectados e não infectados pelo HIV-1 de São Paulo e 18 a protótipos dos subtipos A a E, mostraram mapas de restrição enzimática característicos dos principais subtipos virais descritos até o momento. Tomando como base apenas as enzimas disponíveis comercialmente, foram selecionadas cinco que se mostraram úteis para a subtipagem de HHV-8: Taq I, Nsi I, Hinf I, Hae III e Mse I. Os resultados obtidos com a técnica de PCR-RFLP (reação em cadeia de polimerase associada à análise do polimorfismo de fragmentos de restrição enzimática) mostraram que de 48 espécimes brasileiros previamente classificados como sendo dos subtipos A, B e C por seqüenciamento gênico, todos foram corretamente subtipados pela técnica de PCR-RFLP. Três amostras (duas do subtipo A e uma do B) apresentaram mais um sítio de restrição enzimática além dos descritos como sendo os predominantes. Mais recentemente, outras 27 amostras de 18 casos de infecção por HHV-8 foram subtipadas pela PCR-RFLP. Houve detecção de oito isolados do subtipo A, sendo seis de variante predominante, um de variante minoritária conhecida e um de nova variante viral. Dois casos de infecção por HHV-8 do subtipo B e sete do subtipo C também foram identificados. Finalmente, um provável caso de infecção pelo subtipo E foi encontrado em paciente com SK-AIDS disseminado, resistente à quimioterapia. Na avaliação global, houve maior número de casos de infecção por HHV-8 dos subtipos A e C. Concluindo, devido à alta sensibilidade e especificidade, baixo custo, rapidez e facilidade de execução, a técnica de PCR-RFLP pode ser usada em larga escala para estudos de epidemiologia molecular, principalmente em países em desenvolvimento. / AIDS epidemic has increased the incidence rates of Kaposi’ s sarcoma (KS) in all countries, and KS has been considered an AIDS-defining illness. Since the discovery of the human herpesvirus 8 (HHV-8), the etiological agent of KS, several studies have been conducted in order to characterize HHV-8 in all forms of KS: classic, endemic, iatrogenic, and epidemic. The HHV-8 genome presents a hypervariable region termed ORF K1 useful for virus subtyping. The objectives of the present study were to describe an alternative method for subtyping HHV-8, to compare this new method with DNA sequencing, and to use this method for HHV-8 subtyping. After cloning and sequencing a segment of the ORF K1 (VR1) in 50 HHV-8/DNA isolates from 36 Brazilian KS-AIDS patients, we searched for restriction enzymatic sites in this segment of DNA, and compared them with 18 sequences reported in the literature. Then we constructed the enzymatic restriction maps useful for discriminating all HHV-8 subtypes described up to now, and standardized a PCR-RFLP (restriction fragment length polymorphism analysis) using five commercial enzymes: Taq I, Nsi I, Hinf I, Hae III e Mse I. After comparing the results obtained by the two methods, we used PCR-RFLP for HHV-8 subtyping in 27 new HHV-8/DNA isolates. The results obtained by DNA sequencing and PCR-RFLP showed 100% of concordance, and allowed the use of PCR-RFLP for HHV-8 subtyping. Indeed, we disclosed that among KS-AIDS patients from São Paulo, subtypes A and C are more prevalent than subtype B. Although additional restriction sites were detected in some Brazilian HHV-8 isolates, the majority of them belonged to the predominant strains described in the literature. Interestingly, one probable case of HHV-8 subtype E was detected in a patient who presented disseminated KS and resistance to chemotherapy. Because of its high sensitivity, specificity, low cost, and rapid execution, PCR-RFLP could be used on a large scale, mostly in countries with poor resources and where KS is endemic.

Mycoplasma (M.) bovis est une bactérie pathogène des bovins, à l'origine de signes cliniques divers, comme des mammites, des arthrites, des otites et des bronchopneumonies, ces dernières étant majoritaires en France. Les mycoplasmoses à M. bovis ont un fort coût économique et leur contrôle impose une importante mobilisation sanitaire et un recours très fréquent à l'antibiothérapie. Peu de données étaient disponibles jusque récemment concernant le typage moléculaire et l'antibiosensibilité des souches françaises de M. bovis. Deux études antérieures à ce travail et réalisées au sein de l'UMR « Mycoplasmoses des ruminants » ont montré que les isolats cliniques de M. bovis collectés en France après 2000 appartiennent à un sous-type moléculaire majoritaire (ST2), très homogène et sont par ailleurs multirésistants à la plupart des familles antibiotiques à l'exception des fluoroquinolones. Ces résultats suggèrent la diffusion sur le territoire national d'un clone unique multirésistant. Le premier objectif de cette étude était de déterminer les mécanismes à la base de la perte de sensibilité aux antibiotiques des isolats français. Dans un deuxième temps, les liens entre les différents sous-types moléculaires, les profils d'antibiosensibilité, les maladies associées et le polymorphisme des gènes cibles des antibiotiques ont été investigués. Cette approche a été déployée pour trois familles d'antibiotiques utilisées en pratique vétérinaire: les macrolides, les tétracyclines et également les fluoroquinolones, quoique récemment classées comme molécules critiques. De façon générale, les mutations identifiées dans les cibles des antibiotiques expliquent à elles seules les phénotypes de résistance observés. Des mutations dans les ARNs ribosomaux, cibles des macrolides et des tétracyclines, ont été observées sur des isolats cliniques dès 1978 et sont devenues systématiques sur tous les isolats collectés après 2000 et appartenant au sous-type ST2 majoritaire. En ce qui concerne les fluoroquinolones, la faible augmentation des CMI (concentrations minimales inhibitrices) mesurée chez la plupart des isolats cliniques récents n'a pas été associée à des mutations des QRDR (« Quinolones Resistance-Determining Regions »). Par contre, des altérations cumulées de façon séquentielle dans ces QRDR, associées à une hausse des CMI, ont été mises en évidence lors d'expériences de sélection in vitro et majoritairement pour des souches appartenant à un sous-type récent minoritaire, ST3, apparemment plus variable et plus apte à fixer les mutations. En 2013, le premier isolat clinique présentant une CMI augmentée aux fluoroquinolones a été isolé: il appartient à ce sous-type ST3. L'ensemble des résultats obtenus montrent que les différents sous-types de M. bovis n'évoluent pas de la même façon vers la résistance. Ce constat ajouté à celui de la multirésistance des isolats récents (ST2 ou ST3) met en exergue l'intérêt de la surveillance (sous-typage et antibiosensibilité) et le suivi de l'évolution des isolats de M. bovis circulant en France. Ce suivi permettrait notamment d'anticiper une éventuelle émergence de la résistance aux fluoroquinolones / Mycoplasma (M.) bovis is a bacterial pathogen for cattle, responsible for various clinical signs, like mastitis, arthritis, otitis and respiratory diseases, the latter being the main syndrome present in France. Mycoplasmoses have a great economic impact and their control entails drastic sanitary measures and a frequent use of antibiotherapy. Few data was available until recently on the molecular subtyping and the antimicrobial susceptibility of the French strains of M. bovis. Two previous studies done in the UMR « Mycoplasmoses des ruminants » proved that clinical isolates collected in France after the year 2000 belonged to one major subtype (ST2), which is very homogeneous, and that they were multiresistant to the main antimicrobial families except fluoroquinolones. These results suggested the diffusion of one unique multiresistant clone on the national territory. The first aim of the present study was to decipher the molecular mechanisms underlying the loss of susceptibility to antimicrobials of the French strains. Secondly the links between the molecular subtypes, the antibiotics susceptibility profiles, the clinical origins and the polymorphisms of the target genes were assessed. This approach was used for 3 antimicrobial families currently used in veterinary medicine: macrolides, tetracyclines and fluoroquinolones, although recently classified as critical. Actually, the point mutations observed in the target genes of the antimicrobials accounted for the observed resistance phenotypes. Some mutations in the ribosomal RNAs, targets of the macrolides and the tetracyclines, were observed in clinical isolates as soon as 1978 and they were generalized in all isolates collected after 2000 and belonging to the major subtype ST2. Concerning the fluoroquinolones, the slight increase in MIC (Minimum Inhibitory Concentration) observed in most of the recent isolates was not associated with mutations in the QRDR (Quinolone Resistance-Determining Regions). However alterations that were associated with increased MICs were highlighted and proved to be sequentially cumulated during experiments of in vitro selection under antimicrobials pressure. This was mainly true for strains belonging to a recent and uncommon subtype, ST3, which is apparently more variable and more able to fix the mutations. In 2013 the first clinical strain showing an increased MIC to fluoroquinolones was isolated and proved to belong to ST3. The whole results of this study showed that the different subtypes did not evolve with the same speed towards resistance. This fact, associated with the multiresistant phenotype of the recent isolates (ST2 or ST3), highlights the urge to monitor (subtyping and antimicrobial susceptibility profiles) and to follow-up the evolution of the isolates of M. bovis circulating in France in order to anticipate a potential emergence of the resistance to fluoroquinolones

Cette thèse porte sur l'analyse statique de code par typage dans le but de détecter les erreurs dans les programmes avant leur exécution. Plus précisément, nous nous intéressons ici au domaine du sous-typage, dans lequel les propriétés du code sont représentées par des ensemble de contraintes de la forme (t1 <= t2). Nos mécanismes de vérification sont alors basés sur l'agrégation de contraintes de sous-typage et la vérification de leur compatibilité par saturation. Le langage de base sur lequel nous travaillons est un ML étendu, muni de variants et d'un mécanisme de filtrage de motifs. Nous commençons par définir un formalisme nous permettant d'exprimer nos systèmes de types sous forme de règles d'inférences. Ce formalisme présente l'avantage d'être suffisamment souple pour nous permettre de prouver les propriétés de validité et de terminaison de nos systèmes, et suffisamment précis pour nous permettre d'en dériver une implémentation de manière systématique. Après avoir défini un système de types de base pour notre langage, nous en présentons trois extensions originales : * Une amélioration du typage du filtrage de motifs basée en particulier sur l'ajout d'un opérateur de disjonction entre les contraintes de sous-typage. Cet opérateur permet alors d'exprimer, pour chaque cas de filtrage, le lien entre le filtre et les contraintes extraites du typage de l'expression correspondante. Ceci nous permet en particulier de représenter beaucoup plus finement le type de certaines fonctions et ainsi d'accepter plus de programmes valides. * Une alternative au mécanisme classique de généralisation permettant de distinguer les contraintes associées aux différents usages des paramètres des fonctions. Un tel mécanisme rend en particulier la construction de langage "let" de ML obsolète. Mixé avec la première extension, nous obtenons un système permettant d'encoder dans le langage lui même (c'est à dire sans ajouter de construction supplémentaire), un modèle objet intéressant. * Une formalisation des GADT basée sur une implantation originale des variables de type existentielles. En plus d'être compatible avec le sous-typage, cette variante des GADT présente une amélioration notable par rapport aux GADT standards par le fait qu'elle étend les possibilités d'inférence. Les annotations de type, habituellement obligatoires en présence de GADT, deviennent ici presque toutes facultatives. Bien qu'il soit possible de dériver directement une implémentation de ces systèmes, ce qui est principalement utile pour leur compréhension et leur prototypage, les performances des typeurs obtenus de la sorte ne sont pas suffisantes pour analyser des programmes de taille réelle. Ceci est principalement dû aux différentes extensions que nous apportons au langage des contraintes, en particulier les opérateurs de disjonction et de négation. Nous présentons alors les différentes techniques que nous avons mises en place pour l'implémentation de nos systèmes permettant à nos analyses de passer à l'échelle en pratique. / This PHD thesis focuses on static analysis of programs by type inference in order to detect program errors before their execution. More precisely, we focus hear in the field of sub-typing, where program properties are described by sets of constraints of the form (t1 <= t2). Our verification mechanisms are based on the aggregation of sub-typing constraints and checking of their compatibility by saturation. The base language on which we define our type systems is an ML-like language provided with variants and pattern matching. We starts by defining a formalism to express our type systems thanks to inference rules. This formalism has the advantage to be sufficiently flexible to allow proving validity and termination properties of our systems, and sufficiently precise to allow a systematic derivation of our inference rules into a runnable typer. After the definition of a base type system for our language, we present three novel extensions: * An improvement of type inference for the pattern matching based on the addition of the "or" operator between sub-typing constraints. This operator allow to express a link, in each cases of a match, between the pattern and the constraints generated at typing time of the case expression. This allows us to refine the type of some functions, and then to accept more valid programs. * A new implementation of the generalization mechanism. This allows to distinguish constraints associated to the different occurrences of a function parameter in its body. Thanks to this mechanism, the "let" construction from ML is in particular obsolete. By mixing this extension with the first one, we obtain a type system able to encode "objects" without any additional language construction. * A formalization of GADT based on an novel implementation of existential type variables. In addition to be compatible with the sub-typing context of this thesis, this alternative to GADT has the advantage to improve type inference. As a consequence, most of type annotations, usually required in the presence of GADT, are now optional. Despite the fact that it is possible to directly derive an implementation of our type systems from their rules, that is principally interesting for their comprehension and prototyping, the effectiveness of such typer is insufficient to analyze real world programs. This is principally due to the extensions we provide to the language of constraints, and in particular to the "or" and "not" operators. At then end, we present multiple techniques we used in our implementation to extend the scalability of our analysis.

Classification of lymphoid neoplasms is based mainly on histologic, immunologic, and (rarer) genetic features. It has been supplemented by gene expression profiling (GEP) in the last decade. Despite the considerable success, particularly in associating lymphoma subtypes with specific transcriptional programs and classifier signatures of up- or downregulated genes, competing molecular classifiers were often proposed in the literature by different groups for the same classification tasks to distinguish, e.g., BL versus DLBCL or different DLBCL subtypes. Moreover, rarer sub-entities such as MYC and BCL2 “double hit lymphomas” (DHL), IRF4-rearranged large cell lymphoma (IRF4-LCL), and Burkitt-like lymphomas with 11q aberration pattern (mnBLL-11q) attracted interest while their relatedness regarding the major classes is still unclear in many respects. We explored the transcriptional landscape of 873 lymphomas referring to a wide spectrum of subtypes by applying self-organizing maps (SOM) machine learning. The landscape reveals a continuum of transcriptional states activated in the different subtypes without clear-cut borderlines between them and preventing their unambiguous classification. These states show striking parallels with single cell gene expression of the active germinal center (GC), which is characterized by the cyclic progression of B-cells. The expression patterns along the GC trajectory are discriminative for distinguishing different lymphoma subtypes. We show that the rare subtypes take intermediate positions between BL, DLBCL, and FL as considered by the 5th edition of the WHO classification of haemato-lymphoid tumors in 2022. Classifier gene signatures extracted from these states as modules of coregulated genes are competitive with literature classifiers. They provide functional-defined classifiers with the option of consenting redundant classifiers from the literature. We discuss alternative classification schemes of different granularity and functional impact as possible avenues toward personalization and improved diagnostics of GC-derived lymphomas

In the realm of biomedical technology, both accuracy and consistency are crucial to the development and deployment of these tools. While accuracy is easy to measure, consistency metrics are not so simple to measure, especially in the scope of biomedicine where prediction consistency can be difficult to achieve. Typically, biomedical datasets contain a significantly larger amount of features compared to the amount of samples, which goes against ordinary data mining practices. As a result, predictive models may fail to find valid pathways for prediction during training on such datasets. This concept is known as underspecification. Underspecification has been more accepted as a concept in recent years, with a handful of recent works exploring underspecification in different applications and a handful of past works experiencing underspecification prior to its declaration. However, underspecification is still under-addressed, to the point where some academics might even claim that it is not a significant problem. With this in mind, this thesis aims to identify and minimize underspecification of deep learning cancer subtype predictors. To address these goals, this work details the development of Predicting Underspecification Monitoring Pipeline (PUMP), a software tool to provide methodology for data analysis, stress testing, and model evaluation. In this context, the hope is that PUMP can be applied to deep learning training such that any user can ensure that their models are able to generalize to new data as best as possible.

No description available.

Les toxines botuliques (BoNTs) sont les substances les plus toxiques connues. Elles sont responsables du botulisme, une maladie rare mais le plus souvent mortelle sans prise en charge médicale. Cependant, les applications médicales des BoNTs sont de plus en plus nombreuses du fait de leurs propriétés paralysantes. Leur toxicité par voie inhalée en fait un des 6 principaux agents du risque intentionnel. Les BoNTs, produites par Clostridium botulinum, se répartissent en 7 types sérologiques qui se déclinent en sous-types. Cette biodiversité rend difficile leur identification par les méthodes classiques utilisées pour les toxines protéiques (approches immunologiques). Jusqu’à présent, seule l’analyse génétique permettait de distinguer les différents sous-types entre eux. Dans ce travail a été développée une méthode d’analyse en LC-QqQ-MS/MS en mode MRM pour identifier les différents sous-types de la BoNT/A dans des matrices complexes à partir de peptides communs et spécifiques à ces sous-types. Un traitement d’échantillon par immunocapture sur billes magnétiques couplées à des anticorps anti-peptides a été développé pour isoler la toxine de l’échantillon avant analyse. Des surnageants de culture des sous-types A1 à A3, A5, A7 à A8 ont été utilisés pour valider la méthode. La limite de détection de la méthode est compatible avec les taux de toxine retrouvés habituellement dans les échantillons naturellement contaminés. Cette méthode de spectrométrie de masse a ensuite été utilisée pour quantifier les différents sous-types de la BoNT/A dans une matrice complexe (surnageants de culture de C. botulinum). Une technique de quantification, utilisant un isotope stable de la chaine légère de type A1, ([13C6]K et [13C6]R), a été retenu comme étalon interne. Les différents sous-types de BoNT/A ont été quantifiés dans les surnageants et la quantité de BoNT correspondante à une dose létale minimale de 100% a été déterminée pour chaque sous-type. / Botulinum neurotoxins (BoNTs) are the most poisonous substances known. They are responsible for human botulism, a rare but potentially fatal disease if not quickly treated. However, BoNTs were approved for the treatment of numerous medical applications due to their temporary paralysis effects. BoNTs are among the six agents with the highest risk of potential use as bio-weapons because of their high toxicity in aerosol form. BoNTs, produced by Clostridium botulinum, are divised into seven toxinotypes and each toxinotype contains several subtypes. This biodiversity makes more difficult their identification with classical methods by immunological ways. Until now, only molecular genetical methods could differenciate subtypes among them. The aim of this work was to develop a liquid chromatography tandem mass spectrometry (LC-MS/MS) in MRM mode to efficiently discrimate the distinct subtypes from specific and common peptides. Immunocapture sample preparation with antipeptides antibodies was used and allowed the isolation of the toxin from the sample. Subtyping was performed with crude supernatants (BoNT/A1 to /A3, /A5, /A7 and /A8) in order to validate the method. Limit of detection (LOD) of the proposed method is in the range of minimal toxin concentration found in naturally contamined samples. In a second part of this work, this mass spectrometry method was used to quantify the neurotoxin in complex matrices (supernatants of Clostridium botulinum cultures). Isotope labeled light chain (13C6]K et [13C6]R) from botulinum A1 neurotoxin was produced and used as internal standart. Subtypes were quantified in supernatants and the quantity of neurotoxin for one minimal lethal dose 100% was determined for each subtype

Actuellement, l’implémentation des services (modèles SOAP et RESTful) et de leurs applications clientes est de plus en plus basée sur la programmation par objet. Ainsi, les cadriciels orientés-objets pour les services Web sont essentiellement composés de deux couches : une couche à objets qui enveloppe une couche à services. Dans ce contexte, deux principes sont nécessaires pour la spécification de ces cadriciels : (i) En premier lieu, un couplage faible entre les deux couches, ce qui permet de cacher la complexité des détails techniques de la couche à services dans la couche à objets et de faire évoluer la couche à services avec un impact minimal sur la couche à objets (ii) En second lieu, une interopérabilité induite par le principe de substitution associée au sous-typage dans la couche à objets. Dans cette thèse, nous présentons d’abord les faiblesses existantes dans les cadriciels orientés-objets liés à ces deux principes. Ensuite, nous proposons une nouvelle spécification pour ces cadriciels en vue de résoudre ces problèmes. Comme application, nous décrivons la mise en œuvre de notre spécification dans le cadriciel cxf, à la fois pour SOAP et RESTful. / Today, the implementation of services (SOAP and RESTful models) and of client applications is increasingly based on object-oriented programming languages. Thus, object-oriented frameworks for Web services are essentially composed with two levels: an object level built over a service level. In this context, two properties could be particularly required in the specification of these frameworks: (i)First a loose coupling between the two levels, which allows the complex technical details of the service level to be hidden at the object level and the service level to be evolved with a minimal impact on the object level, (ii) Second, an interoperability induced by the substitution principle associated to subtyping in the object level, which allows to freely convert a value of a subtype into a supertype. In this thesis, first we present the existing weaknesses of object-oriented frameworks related to these two requirements. Then, we propose a new specification for object-oriented Web service frameworks in order to resolve these problems. As an application, we provide an implementation of our specification in the cxf framework, for both SOAP and RESTful models.