No description available.

Mon travail de thèse a été consacré à l’optimisation et l’utilisation de techniques en protéomique clinique, et plus particulièrement de l’imagerie MALDI, pour la recherche et l’identification de marqueurs pathologiques. C’est pourquoi nous avons, dans un premier temps, développé et appliqué des outils d'analyse statistique basés sur la PCA et la classification hiérarchique, appelés PCA-SDA. Ceux-ci offraient une combinaison intéressante avec l'imagerie MALDI et permettaient une simplification des données, une recherche fine des variations moléculaires au sein du tissu et une classification sur la base des profils moléculaires obtenus localement sur les tissus. Appliquée ensuite sur des études de biopsies de cancer de l’ovaire, cette approche nous a permis de détecter et d’identifier plusieurs marqueurs potentiels jouant un rôle dans la réponse du système immunitaire, l’adhésion et l’invasion tumorale. Or, ces mécanismes sont connus pour impliquer des protéases, les proprotéines convertases, dans la maturation des différentes protéines impliquées dans le développement tumoral. Dans ce contexte, nous avons étudié leur expression dans le cancer de la prostate. Il s’est avéré que seule PACE4 était surexprimée dans cette pathologie et nous avons pu établir son rôle primordial à la fois dans la prolifération cellulaire à l’aide d’études in vitro, et dans l’évasion apoptotique par approche protéomique, suite à l’identification de TRPS1, un facteur de transcription impliqué dans l’apoptose. Le rôle prépondérant de PACE4 fait de cette enzyme une cible thérapeutique potentielle du cancer. / My PhD’s work has been completely devoted to the optimization and the use of technologies in clinical proteomics, especially MALDI imaging, for research and identification of disease markers. Therefore we have initially developed and applied the tools of statistical analysis based on PCA and hierarchical clustering, called PCA-SDA, which offered an interesting combination with MALDI imaging and allow simplification of data, fine search of molecular changes within the tissue and a classification based on molecular profiles obtained locally on the tissues. Then applied on ovarian cancer biopsies study, this approach allowed us to detect and identify several potential markers playing a role in immune response, adhesion and tumor invasion. However, these mechanisms are known to involve proteases, like proprotein-convertases, in the maturation of various proteins implicated in tumor development. In this context, we studied their expression in prostate cancer. It pointed that only PACE4 was over-expressed in this disease and we were able to establish its role in cell proliferation using in vitro analysis and in apoptotic evasion with the identification of TRPS1, a transcription factor involved in apoptosis, by proteomics approach. So, PACE4 is a potential therapeutic target for cancer due to its leading role in tumor cell capacities.

No description available.

Les travaux effectués dans le cadre de cette thèse portent sur l’identification automatique de visages dans des bases de données d’images numériques. L’objectif est de simplifier le déroulement de l’identification biométrique qui consiste à la recherche de l’identité requête parmi l’ensemble des individus de la base de données préenregistrée, appelée aussi galerie. En effet, le schéma d’identification classique est complexe et très coûteux notamment dans le cas de grandes bases de données biométriques. Le processus original que nous proposons ici a pour objectif de réduire la complexité et d’améliorer les performances en terme de temps de calcul et de taux d’identification. Dans ce cadre biométrique, nous avons proposé une étape de classification non supervisée ou clustering des visages pour partitionner la base d’images en plusieurs sous ensembles cohérents et bien discriminés. Pour ce faire, l’algorithme de clustering vise à extraire, pour chaque visage, un ensemble de descripteurs, appelé signature, qui lui soit spécifique. Trois techniques de représentation faciales ont été développées dans le but d’extraire des informations différentes et complémentaires décrivant le visage humain : deux méthodes factorielles d’analyse multidimensionnelle et de projection des données (méthodes dites « Eigenfaces » et « Fisherfaces ») ainsi qu’une méthode d’extraction des moments géométriques de Zernike. Sur la base des différentes classes de signatures extraites pour chaque visage plusieurs méthodes de clustering sont mises en compétition afin de dégager la classification optimale qui conduit à une meilleure réduction de la galerie. Les méthodes retenues sont, soit de type « centres mobiles » telles que les K-moyennes et les centres mobiles de Forgy, ou de type « agglomérative » telle que la méthode de BIRCH. Sur la base de la dépendance des partitions générées, ces différentes stratégies classificatoires sont ensuite combinées suivant une architecture parallèle de manière à optimiser la réduction de l’espace de recherche à un sous ensemble de la base de données. Les clusters retenus in fine étant ceux pour lesquels la probabilité d’appartenance de l’identité requête est quasi certaine. / The work done in the framework of this thesis deal with the automatic faces identification in databases of digital images. The goal is to simplify biometric identification process that is seeking the query identity among all identities enrolled in the database, also called gallery. Indeed, the classical identification scheme is complex and requires large computational time especially in the case of large biometric databases. The original process that we propose here aims to reduce the complexity and to improve the computing time and the identification rate performances. In this biometric context, we proposed an unsupervised classification or clustering of facial images in order to partition the enrolled database into several coherent and well discriminated subsets. In fact, the clustering algorithm aims to extract, for each face, a specific set of descriptors, called signature. Three facial representation techniques have been developed in order to extract different and complementary information which describe the human face: two factorial methods of multidimensional analysis and data projection (namely called "Eigenfaces" and "Fisherfaces") and a method of extracting geometric Zernike moments. On the basis of the different signatures obtained for each face, several clustering methods are used in competing way in order to achieve the optimal classification which leads to a greater reduction of the gallery. We used either "mobile centers" methods type such as the K-means algorithm of MacQueen and that of Forgy, and the "agglomerative" method of BIRCH. Based on the dependency of the generated partitions, these different classifying strategies are then combined using a parallel architecture in order to maximize the reduction of the search space to the smallest subset of the database. The retained clusters in fine are those which contain the query identity with an almost certain probability.

Monophyly of the gomphoid-phalloid clade was confirmed based on multigene phylogenetic analyses. Four major subclades (Hysterangiales, Geastrales, Gomphales and Phallales) were also demonstrated to be monophyletic. The interrelationships among the subclades were, however, not resolved, and alternative topologies could not be rejected statistically. Nonetheless, most analyses showed that the Hysterangiales and Phallales do not form a monophyletic group, which is in contrast to traditional taxonomy. The higher-level phylogeny of the gomphoid-phalloid fungi tends to suggest that the Gomphales form a sister group with either the Hysterangiales or Phallales. Unweighted parsimony character state reconstruction favors the independent gain of the ballistosporic mechanism in the Gomphales, but the alternative scenario of multiple losses of ballistospory could not be rejected statistically under likelihood based reconstructions. This latter hypothesis is consistent with the widely accepted hypothesis that the loss of ballistospory is irreversible. The transformation of fruiting body forms from nongastroid to gastroid was apparent in the lineage leading to Gautieria (Gomphales), but the tree topology and character state reconstructions supported that truffle-like taxa of the Phallales are ancestral to stinkhorns, which possess more complex, epigeous fruiting bodies. Importantly all taxa within the Phallales are statismosporic and thus the derived stinkhorn morphology does not require an independent gain of ballistospory. Biogeographical analyses of the Hysterangiales strongly suggest that the ectomycorrhizal lineages within the Hysterangiales originated in the East Gondwana. The synonymous substitution rate indicated a Paleozoic origin of the Hysterangiales although a possibility of a Cretaceous origin could not be discarded. Because modern ectomycorrhizal plants were absent during the Paleozoic era, a potential existence of the Hysterangiales during this time must be explained either by novel ectomycorrhizal association of the Hysterangiales with unknown plant lineages, or multiple, independent gains of ectomycorrhizal habit. The Paleozoic origin of the Hysterangiales also indicates that mycophagous animals may not be the most important factor for range expansions of the Hysterangiales. Taxonomic revisions are made for the gomphoid-phalloid fungi. One subclass (Phallomycetidae), two orders (Hysterangiales and Geastrales), four families (Gallaceaceae, Phallogastraceae, Trappeaceae and Sclerogastraceae), 7 genera (Austrohysterangium, Cribbangium, Rodwayomyces, Beeveromyces, Cazomyces, Insulomyces and Viridigautieria) and 22 new combinations are proposed. / Graduation date: 2006

Graduation date: 1985

Graduation date: 1996

Infrageneric classification in Ribes has previously relied on a few, often conflicting, morphological markers, such as spines, glands, and inflorescence morphology. Suggestions that hybridization drives the evolution of the genus have not been tested using phylogenetic methods. To assess the validity of infrageneric classifications and the importance of hybridization to the evolution of the group, and to develop an explicit phylogenetic hypothesis, I surveyed exemplars from all subgenera for restriction site variation in two cpDNA regions. Parsimony analysis shows that red currants, European alpine currants, golden currants, true gooseberries, and western gooseberries appear on separate clades. A less well supported Glade includes the western North American ornamental currant group and a portion of the dwarf currants. The presence of distinct lineages in Ribes is strongly supported by characters with very high consistency, suggesting that hybridization among infrageneric groups is not common in the genus. Unexpectedly, spiny currants and true gooseberries are united, suggesting either a sister group relationship or the possibility that one of these groups arose by reticulate evolution. The four black currant species examined exhibit surprisingly high divergence, and are not monophyletic in the analysis. Maximum likelihood analysis supports these results. Basal relationships of these lineages are not well resolved. A similar analysis of a portion of the nuclear ribosomal repeat produced very few characters. Although these data are highly homoplasious, their analysis bears some resemblance to that of the chloroplast DNA. Most prominently, the ornamental currant Glade has identical membership. Brief reviews of the scattered palynological and paleontological literature concerning Ribes are presented. / Graduation date: 1995

The taxonomic configuration of the Order Primates in three different schemes (morphological, immunological, and cladistic) is examined. Multivariate techniques of analysis are used to testthe taxonomies, one against the others, resulting in a quantitative comparison. The analysis reveals both the similarities and differences among the taxonomies, resulting ultimately in a composite classification of the Order Primates. The composite taxonomy is a quantitatively meaningful and qualitatively interpretable result.

Thesis (PhD)(Botany))-Universiteit van Pretoria, 1992. / Summary in English and Afrikaans. Includes bibliographical references.