La découverte de motifs fréquents est un des problèmes en fouille de données. Afin de mieux comprendre l'influence des données sur les algorithmes, nous présentons une étude expérimentale des jeux de données communément utilisés par la communauté. Cette étude permet d'aboutir à une nouvelle classification des données en fonction des bordures : stable et en accord avec les performances des algorithmes. Malgré le grand nombre de travaux et un cadre théorique des problèmes d'extraction de motifs intéressants, l'utilisation de ces algorithmes pour résoudre des problèmes "équivalents" est peu répandue et reste délicate. Face à ces limites, nous proposons un algorithme générique de découverte des bordures des motifs intéressants, appelé ABS (Adaptive borders Search), adaptant dynamiquement sa stratégie en fonction des données. De plus, une librairie générique de composants C++ a été proposée pour faciliter le développement de solutions logicielles pour cette famille de problèmes.

The vampire figure is very much a part of the literary landscape of today, and has been so for the last 200 years. The vampire has not always appeared as it does today, as the rich, urbane gentleman, but has its origins in old folklore legends. The idea that the vampire figure has changed over the course of history is not new, but instead of discussing the phenomena influencing, and changing, the vampire motif, this essay will try to shed light on the aspects of the folklore vampire that are still part of the vampire of today. By applying the theory of folklorism (folklore not in its original context, but rather the imitation of popular themes by another social class, or the creation of folklore for purposes outside the established tradition), presented by Hans Moser and Hermann Bausinger among others, this essay attempts to prove that the modern vampire is in fact a folklorism of the old folklore legends. The essay examines the more recent incarnation of the vampire, the literary vampire who emerged during the 18th and 19th century, with the intent to prove that, while it is different from its origin, it has several features in common with its ancestry as well. To show this, examples from Bram Stoker’s Dracula (1897), and the more recent novel Fangland (2007) by John Marks have been chosen to serve as basis for the analysis. Both novels clearly show instances where folklore has been brought into the narrative as a way to define and depict the vampire.

The following thesis research was undertaken to gain a better understanding of the mechanisms that regulate the cellular trafficking and signaling of the endothelial differentiation gene (EDG) family of G-protein coupled receptors, LPA1, LPA2, and LPA3. This thesis will specifically focus on the regulation of the trafficking of the LPA1 Lysophosphatidic acid receptor, which is the most widely expressed and has been shown to be a major regulator of migration of cells expressing it. The initial studies undertaken in this project were aimed at understanding the endocytic pathway followed by the LPA1 receptor. Lysophosphatidic acid (LPA), an abundant serum phospholipid, stimulates heterotrimeric G protein signaling by activating three closely related receptors, termed LPA1, LPA2 and LPA3. In the first part of the project we show that in addition to promoting LPA1 signaling, membrane cholesterol is essential for the association of LPA1 with β-arrestin, which leads to signal attenuation and clathrin dependent endocytosis of LPA1. The second phase of the project was aimed at elucidating the different structural motifs required for the trafficking and signaling of the LPA1 receptor and helping us gain a more mechanistic view of the processes involved in its regulation. In the second part of the project we show that agonist-independent internalization of the LPA1 receptor is clathrin adaptor, AP-2 dependent and PKC-dependent and that it requires a distal dileucine motif, whereas agonist-dependent internalization of the LPA1 receptor is β-arrestin and clathrin-dependent and requires a cluster of serine residues in the tail region, which is upstream of the dileucine motif. These studies collectively vastly enhance our understanding of mechanisms that regulate LPA1 trafficking and signaling. These studies can also be applied to other G-protein coupled receptors making the task easier for other scientists to understand this vast family of receptors.

In this study, the focus is on predicting the subcellular localization of a protein, since subcellular localization is helpful in understanding a protein&rsquo / s functions. Function of a protein may be estimated from its sequence. Motifs or conserved subsequences are strong indicators of function. In a given sample set of protein sequences known to perform the same function, a certain subsequence or group of subsequences should be common / that is, occurrence (frequency) of common subsequences should be high. Our idea is to find the common subsequences through clustering and use these common groups (implicit motifs) to classify proteins. To calculate the distance between two subsequences, traditional string edit distance is modified so that only replacement is allowed and the cost of replacement is related to an amino acid substitution matrix. Based on the modified string edit distance, spectral clustering embeds the subsequences into some transformed space for which the clustering problem is expected to become easier to solve. For a given protein sequence, distribution of its subsequences over the clusters is the feature vector which is subsequently fed to a classifier. The most important aspect if this approach is the use of spectral clustering based on modified string edit distance.

Cette thèse propose un cadre d'évaluation pour des requêtes XQuery dans un<br />contexte de médiation de données XML. Un médiateur doit fédérer des sources de données<br />distribuées et hétérogènes. A cette fin, un modèle de représentation des requêtes est néces-<br />saire. Ce modèle doit intégrer les problèmes de médiation et permettre de définir un cadre<br />d'optimisation pour améliorer les performances. Le modèle des motifs d'arbre est souvent<br />utilisé pour représenter les requêtes XQuery, mais il ne reconnaît pas toutes les spécifica-<br />tions du langage. La complexité du langage XQuery fait qu'aucun modèle de représentation<br />complet n'a été proposé pour reconna^³tre toutes les spécifications. Ainsi, nous proposons un<br />nouveau modèle de représentation pour toutes les requêtes XQuery non typées que nous appe-<br />lons TGV. Avant de modéliser une requête, une étape de canonisation permet de produire une<br />forme canonique pour ces requêtes, facilitant l'étape de traduction vers le modèle TGV. Ce<br />modèle prend en compte le contexte de médiation et facilite l'étape d'optimisation. Les TGV<br />définis sous forme de Types Abstraits de Données facilitent l'intégration du modèle dans tout<br />système en fonction du modèle de données. De plus, une algèbre d'évaluation est définie pour<br />les TGV. Grâce µa l'intégration d'annotations et d'un cadre pour règles de transformation, un<br />optimiseur extensible manipule les TGV. Celui-ci repose sur des règles transformations, un<br />modèle de coût générique et une stratégie de recherche. Les TGV et l'optimiseur extensible<br />sont intégrés dans le médiateur XLive, développé au laboratoire PRiSM.

Thesis (Ph. D.)--University of Washington, 2000. / Vita. Includes bibliographical references (leaves 53-58).

Texte remanié de: Thèse d'état--Histoire moderne--Paris 1, 2001. Titre de soutenance : Décors de papier : production, commercialisation et usages des papiers peints à Paris, 1750-1820. / En appendice, choix de documents. Bibliogr. p. 157-158. Index.

Analysing ecological communities as complex networks of interactions has become an important tool for ecologists. Understanding how these networks change through time, over landscapes, or in response to disturbances is a primary goal of community ecology. The number of interactions and the way in which those interactions organise themselves as individuals, small groups, and the whole community can play an important role in predicting how ecological communities will respond to disturbances. In this thesis, we investigated variation in network structure at several scales both empirically and in a theoretical context. Our first hypothesis was that the structural role of species in a variable system would show little variation, despite high levels of species turnover and a fragmented landscape. In a collaboration with Riikkaa Kaartinen and Tomas Roslin, we studied the distribution of species’ roles at three scales in host-parasitoid networks collected from a fragmented forest in Finland. We found that species’ roles were remarkably consistent through time and in the presence of species turnover. These results suggest that species’ roles may be an intrinsic property of species and may be predictable over spatial and temporal scales. Our second study investigated the structural variation of simulated ecological networks and the relationship between structural variation and whole-network measures of network organization, such as connectance, nestedness, and modularity. We quantified structural variation of networks at three scales, macro-scale, motif-scale, and participation scale. These scales represent whole-network measures (macro-scale), sub-network measures (motifs – small groups of interacting species), and individual measures (motif participation). We compared the variation in these structures to connectance, nestedness, and modularity. We found that at fixed levels of connectance, nestedness, and modularity, the motif profiles of networks and the distribution of species across those profiles showed remarkable dissimilarity. This result suggests that networks displaying similar macro-scale structural measures can be composed of vastly different motif- and participation-scale structures. Together, the work that makes up this thesis suggests that we should give more attention to the meso-scale structures of ecological networks. As the more detailed perspective of motifs can capture additional detail about the structure of empirical networks, and as a result, provide a clearer picture of ecological communities. In addition, we found that the particular species themselves can have a significant impact on the meso-scale structure and, in some cases, may impose strict limitations on what interactions can occur within a community. This has important implications for our understanding of how ecological networks are built and maintained, and thereby for our understanding of the stability and resilience of ecological communities.

published_or_final_version / abstract / toc / Anatomy / Master / Master of Philosophy

published_or_final_version / abstract / toc / Anatomy / Master / Master of Philosophy