25655 research outputs found
Sort by
Model-based Management of Information System Security Risk
No full textDuring the last twenty years, the impact of security concerns on the development and exploitation of information systems never ceased to grow. Security risk management methods are methodological tools, helping organisations to take rational decisions, regarding the security of their IS. Feedbacks on the use of such approaches show that they considerably reduce losses originating from security problems. Today, these methods are generally built around a well structured process. However, the product coming from the different risk management steps is still largely informal, and often not analytical enough. This lack of formality hinders the automation of the management of risk-related information. Another drawback of current methods is that they are generally designed for being used a posteriori, that is, to assess the way existing systems handle risks, and are with difficulty usable a priori, during information system development. Finally, with method using its own terminology, it is difficult to combine several methods, in the aim of taking advantage of each of them. For tackling the preceding problems, this thesis proposes a model-based approach for risk management, applicable from the early phases of information system development. This approach relies on a study of the domain's own concepts.
This scientific approach is composed of three successive steps. The first step aims at defining a reference conceptual model for security risk management. The research method followed proposes to base the model on an extensive study of the literature. The different risk management and/or security standards, a set of methods representative of the current state of the practice, and the scientific works related to the domain, are analysed. The result is a semantic alignment table of the security risk management concepts, highlighting the key concepts taking place in such an approach. Based on this set of concepts, the security risk management domain model is built. This model is challenged by domain experts in standardisation, risk management practitioners and scientists.
The second step of this research work enriches the domain model with the different metrics used in a risk management method. The proposed approach combines two methods to define this set of metrics. The first one is the Goal-Question-Metric (GQM) method applied on the domain model. This method allows to focus on reaching the best return on security investment. The second one enriches the metrics identified with the first approach, through a study of the literature based on standards and methods addressed during the first step. An experimentation on a real case of these metrics is performed, in the frame of supporting a SME towards the ISO/IEC 27001 certification.
Finally, in a third step, a set of conceptual modelling languages dedicated to information security is noticed in the literature. These languages are mainly coming from the requirements engineering domain. They allow to tackle security during the early phases of information system development. The conceptual support proposed by each of them is evaluated, and thus the gap to bridge for being able to completely model the different steps of risk management too. This work ends in an extension proposal of the Secure Tropos language, and a process to follow for using this extension in the frame of risk management, illustrated by an example.(DOCSC06)--FUNDP, 200
Characterisation of the fumarase FumC and of its interaction with the essential histidine kinase PdhS at the old pole of Brucella abortus
No full textRecently the bacterial pathogen Brucella abortus has been shown to asymmetrically divide in a reminiscent way compared to the leguminous symbiont Sinorhizobium meliloti and to the differentiation paradigm Caulobacter crescentus, two bacteria phylogenetically related to B. abortus. In C. crescentus, several proteins belonging to the well-known two-components signal transduction systems take a large part in the establishment of cellular asymmetry and in the differentiation process undergone by the flagellated cell type to become a stalked cell type. Strikingly many of these proteins have a homolog in B. abortus and S. meliloti, as predicted from genomes analysis, which suggests the possibility that a conserved network could operate an equivalent differentiation event in these two bacteria. Furthermore these two host-associated bacteria encode in their genome at least one supplementary protein protagonist.
Therefore we focused our work on a main divergence in B. abortus and C. crescentus signalling network components namely the histidine kinase PdhS. We first demonstrated that pdhS gene is essential and that its protein product is asymmetrically distributed to focus at the old pole of B. abortus. This molecular behaviour strengthens the differentiation hypothesis in B. abortus. We subsequently identified the class II fumarate hydratase FumC (one of the key enzymes in the tricarboxylic acids cycle) as a partner of PdhS that specifically interacts with the "sensory" domain of PdhS. Logically FumC colocalizes with PdhS at the old pole of B. abortus while the B. abortus class I fumarate hydratase FumA that does not interact with PdhS remains scattered throughout the cytoplasm. Surprisingly we also discovered that the positional information encompassed in FumC is altered in its homologs in C. crescentus and S. meliloti. To sum up, our data suggest that the aforementioned bacteria that belong to the same taxonomic class share the common characteristic to differentiate albeit they have developed diverse mechanistic specificities along the evolution.(DOCSC03)--FUNDP, 200
Multilevel Optimization: Convergence Theory, Algorithms and Application to Derivative-Free Optimization
No full textWe present new developments in the context of multilevel trust-region methods
for nonlinear optimization. Motivated by the results obtained for unconstrained problems,
we have extended the convergence theory for bound-constrained problems and for the use of
infinity-norm trust regions. As an alternative for these methods, we have developed an algo-
rithm that uses multilevel techniques for the exact resolution of the trust-region subproblem.
This new method guarantees the convergence of the trust-region algorithm to a second-order
critical point, with a much reduced associated cost when compared to classical methods. Un-
fortunately, there are problems for which we cannot compute the derivatives of the objective
function. The methods used today to solve this kind of problems are limited by the cost of
the objective function computation. We present a multilevel version of one of these methods,
that allows for the treatment of larger instances of the problem, as well as numerical results
obtained with this new algorithm.Nous présentons des nouveaux développements dans le cadre des méthodes de ré-
gion de confiance multi-niveaux pour l’optimisation non-linéaire. Motivés par les résultats
obtenus pour l’optimisation sans contraintes, nous avons développé une théorie de conver-
gence pour les problèmes aux contraintes de borne, et pour les régions de confiance définies
par la norme infinie. Comme alternative à ce genre de méthodes, nous avons développé un al-
gorithme qui utilise des techniques multi-niveaux pour la résolution exacte du sous-problème
de la région de confiance. Cette nouvelle méthode garantit la convergence de l’algorithme à
un point critique du deuxième ordre, avec un coût associé nettement inférieur comparé aux
méthodes classiques. Malheureusement, il y a des problèmes où les dérivées de la fonction
objectif ne peuvent pas être calculées. Les méthodes utilisées pour résoudre ce genre de
problèmes sont limitées par le coût du calcul de la fonction objective. Nous présentons donc
une version multi-niveaux de cette méthode qui permet de traiter des problèmes de taille plus
conséquente, ainsi que des résultats numériques obtenus avec ce nouvel algorithme.(DOCSC00 ) -- FUNDP, 200
Three essays on the economics of aidAn analysis of social labeling and in-kind transfers
No full textThis work proposes a theoretical analysis of two specific forms of aid to developing countries: social labelling and in-kind transfers. Social labels have developed rapidly over the recent years. In 2006, consumers worldwide bought 1,6 billion Euros worth of fair-trade certified products, 42 % more than the year before. Labels are particularly attractive as an instrument for concerned consumers in the North to compensate Southern producers for the cost of complying with some minimum labour requirements. In-kind transfers are one of the most common form of aid and account for a large part of the total Official Development Assistance given.
In the first essay, a model is developed to investigate the impact of a label certifying the absence of child labour in the export production of the South. When most eligible producers in the South can obtain the label, its impact is considerably reduced by a displacement effect whereby adult workers replace children in the export sector while children replace adults in the domestic sector. The label is then unable to create a price differential between goods produced under the label and those produced without it. When only a small fraction of eligible producers have access to the label, so that the South exports both labelled and unlabelled production to the North, labelled producers generally gain while those without a label generally loose from the introduction of the label. Ex ante welfare may thus fall in the South if the probability of getting a label when one qualifies is small. The impact on child labour is in general ambiguous.
The second essay investigates the impact of a label certifying high labour standards in the export production of the South. When the price premium from the North just covers the cost of adopting high labour standards in the South, it is shown that the welfare of Northern consumers increases iff the welfare of Southern producers decreases. Moreover, a label is also not Pareto-improving when only a small fraction of producers have access to the label, so that the South exports both labelled and unlabelled production to the North. When adopting high labour standards is not costly for producers, so that the label resembles to label certifying a wage premium, a label that rises the demand for Southern products is Pareto-improving.
In the third essay, it is shown that one-side altruism can provide a rationale for over-providing in-kind transfers. In the model, a selfish recipient has an incentive to under supply effort and capital in order to manipulate the post-production transfers made by an altruistic donor. When effort is not enforceable by the donor, the donor's best response to this Samaritan's dilemma is to over-provide the recipient with a capital transfer in the pre-production period. This allows the donor to mitigate the dilemma, but it automatically creates an inefficiency since too much capital is invested. In the case the donor cannot tax the recipient, so that at equilibrium the donor pre-commits not to make a post-production transfer by over-providing the recipient with a pre-production transfer, transfers given fully in-cash would lead to an efficient outcome. A capital transfer is however chosen by the donor as it allows to reduce the total amount transferred to the recipient.(DOCSESG00)--FUNDP, 200
Is HYAL-1 an authentic lysosomal enzyme?
No full textIntroduction: Hyaluronan is a glucosaminoglycan (GAG) found in the extracellular matrix (ECM) of many tissues where it influences many biological processes such as cell migration and proliferation. Its levels under normal cellular conditions are held in check by the balance between its synthesis by HA synthases and its catabolism by a family of depolymerising enzymes, Hyaluronidases, six of which are already known in mammals (Csoka et al., 2001). The most predominant and active of these are Hyal-1 and Hyal-2. They degrade the polymer to progressively smaller and different sizes of fragments which exert a wide and occasionally opposing spectrum of biological activities (Stern, 2008). Hyaluronidase-1 is the most ubiquitously distributed hyaluronidase, by virtue of its mRNA expression, in all mammalian tissues with exception of adult brain. The highest specific activity of the protein is found in serum, from which it derives the name, "plasma Hyal-1" (Frost et al., 1997), but it is tightly regulated by association with its inhibitors (Stern, 2005). At the same time, Hyal-1 is the most abundant hyaluronidase in somatic tissues particularly those with the highest turnover rate for HA, its substrate, namely the liver, kidney and the spleen (Csoka et al., 2001; Laurent and Fraser, 1992). Apart from its acidic optimal pH, no other evidence has ever been experimentally established in support of the lysosomal localization that is attributed to the protein. We investigated the subcellular localization of Hyal-1 by classical centrifugation methods, based on comparing the behavior of an intracellular compound under different centrifugation systems with that of known reference marker enzymes under the same conditions.
Methods: Experiments were carried out on perfused rodent liver and kidney as well as on human hepatoma cells, HepG2. Detection of the protein was by immunodetection (Western blotting) and zymography activity. These assays were carried out on perfused tissue fractions obtained after differential centrifugation in isoosmotic sucrose solution, according to the scheme of de Duve et al. (1955). Likewise detection was performed on sucrose and Nycodenz density gradients prepared from M, L and P fractions after isopycnic centrifugation. Reference enzymes for different organelles were assayed parallel to these detection methods so as to establish the relative distribution profile of Hyal-1, on the principle that the intensity of signal in the Hyal-1 blot or zymogram would be comparable to the relative specific activity of the marker enzymes in the same fractions.
Results: the distribution profile of Hyal-1in blot differed from that of â-galactosidase (lysosomes) in both HepG2 cells and rodent liver. However, it was realized, on testing the antibody used -1D10, with preparation from Hyal-1 -/- mouse liver, that the antibody was not specific. It recognized, in addition to Hyal-1, a contaminant molecule whose behavior could be reminiscent to peroxisomal nature. A non-lysosomal localization of the enzyme was suggested but a specific, non-ambiguous detection technique was sought.
Zymography was the choice, supported by total negative results on preparations from both liver and kidney of Hyal-1 -/- mice, thus validating the specificity of the technique for Hyal-1 detection. Differential centrifugation indicated that Hyal-1 was recovered in L and P fractions while it was almost absent in M, fraction in which the bulk of lysosomes are recovered. Isopycnic centrifugation of L and P fractions in sucrose gradients when the animals had received a prior injection of Triton WR 1339, non ionic detergent known to modify lysosome density, revealed that the induced density shift of lysosomes did not affect the distribution of Hyal-1. Furthermore, Hyal-1 remained associated with sedimentable structures when the same fractions were subjected to hypotonic shock contrary to lysosomal enzymes release in the supernatant. Similar results were obtained from both mouse and rat liver and so were results from differential centrifugation of rat kidney.
Having been suggested that Hyal-1 in kidney and liver could be originating from serum Hyal-1, the endocytosis of Hyal-1 was followed by injecting the protein in WT and Hyal-1 -/- mice and comparing its distribution with that of endogenous Hyal-1 enzyme. The injection of rh-Hyal-1(recombinant human Hyal-1) labelled with Iodine 125 in WT mice, unlabelled rh-Hyal-1 injected in Hyal-1 -/- mice and serum of a WT mouse in Hyal-1 -/- mice were carried out. The intracellular journey of endocytosed Hyal-1 was followed by analyzing the distribution of the protein in fractions isolated from homogenates prepared from perfused livers of animals injected at increasing times after injection and fractionated by different methods. The distribution of endocytosed Hyal-1 was monitored by determining the radioactivity or by zymography which were compared with that of marker enzymes, â-galactosidase (lysosomes) and alkaline phosphodiesterase (plasma membrane). Results showed that, Hyal-1 is endocytosed and rapidly transported inside liver cells with a maximum uptake being attained a few minutes after injection then followed by a relatively rapid inactivation. Distribution patterns after both differential and isopycnic centrifugation indicate that probably the protein is endocytosed by sinusoidal cells and that on reaching lysosomes, the protein is rapidly inactivated in the presence of lysosomal hydrolases. The same observation was made with in vitro incubation of endogenous and rh-Hyal-1 with purified lysosomes.
Conclusion: Our results agree with the proposal that tissue Hyal-1 originates from endocytosed serum enzyme. The intracellular distribution of the enzyme, on the other hand, would be determined by the steady state situation in which the constant concentration of the Hyal-1 in serum and the relatively rapid inactivation of the protein on reaching terminal lysosomes, site of the bulk of lysosomal hydrolases. Conceivably, with the concentration of the endocytosed Hyal-1 remaining high enough in the endosomal compartments, partial digestion of the endocytosed HA would probably be realized in the prelysosomal compartments, prior to inactivation of Hyal-1 on being transferred to terminal lysosomes.(DOCMED00)--FUNDP, 200
Le problème de la définition des entités linguistiques chez Ferdinand de Saussure
No full textThe question this essay seeks to answer can be formulated in a seemingly simple manner, namely: if language is a “system”, what elements are its constituents? This simplicity, however, is fallacious and hides a deep intricacy. An acceptable answer to this question would in fact not simply lie in an affirmation which states, for example, that elements of ‘Language’ system are such: it should also incorporate an explanation of how these elements interact, a definition of their intrinsic properties and some precision about their common characteristics. An explanation of all that permits one to invoke the notion of “elements” as part of a “system”, and about a "system" composed of that kind of “elements”. The description of an “element” is strictly equivalent – as Saussure taught – to a description of the system in which this element is involved, that is to a determination of (types of) relationships linking system’s elements. Viewed in this light, the question about what are the entities involved in language system is an issue that bears directly and simply on the Language system’s concept, as Saussure conceived.
This thesis has three parts. The first one, devoted to Saussurean concept of “system”, attempts to show that there are variations in Saussure’s development, and that is possible to identify at least two different configurations (one appointed by Saussure “system of oppositions”, the other called “grammatical system”). The second part, devoted to the concept of “value”, tries to show that it is possible to find, in Saussure’s manuscript, at least two different configurations of this concept. One of them would be coherent with purely negative and differential criteria. The other, more complex, would involve elements not able to be reduced to ‘pure differential’ criteria. Our hypothesis (developed in the third part) is to stand that these patterns were developed by Saussure from treatment of different issues, including, consequently, elements definable in different ways: on one hand, "phonemes"; on the other, “signs”.La question qui est au cœur de cette thèse peut donc être formulée d’une manière apparemment simple, à savoir : si la langue est un système, quels sont les éléments qui la constituent ? Cette simplicité n’est pourtant qu’apparente, et dissimule, en réalité, une grande complexité. Une réponse acceptable à cette question consisterait en effet non simplement en une affirmation qui précise, par exemple, que les éléments du système Langue sont tels et tels. Elle devrait comporter, également, une explication de leur mode (ou leurs modes) d’interaction, une formulation de leurs lois, une définition de leurs propriétés intrinsèques et de leurs caractéristiques communes, bref une explicitation de tout ce qui justifie que l’on soit autorisé à parler d’« éléments » faisant partie d’un « système » (en l’occurrence, d’une langue), et d’un « système » composé de ce(s) type(s) d’« éléments ». La description d’un élément équivaut – c’est Saussure qui l’a enseigné – à une description du système auquel cet élément participe, c’est-à-dire à une détermination des (types de) rapports qui relient les éléments entre eux. De ce point de vue, la question de savoir quelles sont les entités qui composent le système Langue est une problématique qui porte de manière directe sur la notion de « système Langue », tout court, tel que Saussure le concevait.
Cette thèse comporte trois parties. La première, consacrée à la notion de « système », essaye de montrer qu’il existe chez Saussure des variations, et qu’il est possible de dégager au moins deux configuration nettement différentes (l’une nommée par Saussure « système d’oppositions », l’autre « système » ou « mécanisme », ou encore « organisme grammatical »).
La deuxième partie, consacrée à la notion de « valeur », tente de montrer qu’il est possible de trouver, chez Saussure, au moins deux configurations différentes. L’une suivant une voie purement négative et différentielle, l’autre, plus complexe, comportant des éléments non réductibles à des différences pures.
Notre hypothèse a été que ces configurations théoriques distinctes sont issues, chez Saussure, du traitement de problématiques différentes, comportant des éléments définissables, par conséquent, de manière différente : d’un côté, des « phonèmes » ; de l’autre, des « signes », en tant qu’entités « doubles », composées d’un signifiant et d’un signifié.(DOCFILO01) -- FUNDP, 200
Program Analysis and Transformation for Data-Intensive System Evolution
No full textData-intensive software systems are generally made of a database (sometimes in the form of a set of files) and a collection of application programs in strong interaction with the former. They constitute critical assets in most enterprises, since they support business activities in all production and management domains. Data-intensive systems form most of the so-called legacy systems: they typically are one or more decade old, they are very large, heterogeneous and highly complex. Many of them significantly resist modifications and change due to the lack of documentation, to the use of aging technologies and to inflexible architectures. Therefore, the evolution of data-intensive systems clearly calls for automated support.
This thesis particularly explores the use of automated program analysis and transformation techniques in support to the evolution of the database component of the system. The program analysis techniques aim to ease the database evolution process, by helping the developers to understand the data structures that are to be changed, despite the lack of precise and up-to-date documentation. The objective of the program transformation techniques is to support the adaptation of the application programs to the new database. This adaptation process is studied in the context of two realistic database evolution scenarios, namely database platform migration and database schema refactoring.Les systèmes d’information sont généralement composés d’une base de données
(parfois sous la forme d’un ensemble de fichiers) et d’une collection de programmes
d’application en forte interaction avec celle-ci. Ces systèmes constituent des composants
critiques dans la plupart des entreprises et organisations, car ils supportent
leurs activités dans tous les domaines de production et de gestion. Les systèmes
d’information forment souvent ce que l’on appelle des systèmes hérités: ils ont été
d´eveloppés il y a plus de dix ans, sont très volumineux, hétérogènes et hautement
complexes. La plupart d’entre eux résistent fortement au changement, de part le
manque de documentation, l’utilisation de technologies obsolètes et d’architectures
peu flexibles. C’est pourquoi l’évolution des systèmes d’information nécessite un
plus grand support automatisé.
Cette thèse se propose d’explorer l’utilisation de techniques d’analyse et de
transformation automatique de programmes, comme support à l’évolution de la
base de données d’un système d’information. Les techniques d’analyse ont pour but
de faciliter le processus d’évolution de la base de données, en aidant les développeurs
à comprendre les structures de données qui doivent évoluer, malgré le manque de
documentation. L’objectif des techniques de transformation de programmes et
de supporter l’adaptation des programmes d’applications à la nouvelle base de
données. Ce processus d’adaptation est étudié dans le contexte de deux scénarios
réalistes d’´evolution: la migration de la base de données vers une nouvelle plateforme
et la restructuration de son schéma.(DOCSC06 ) -- FUNDP, 200
Identification and Characterization of the Interaction between Human Rab2 GTPase and Brucella spp. RicA Protein
No full textBased on a previous high-throughput yeast two-hybrid (Y2H) screen of the Brucella melitensis ORFeome to identify bacterial proteins interacting with human putatively phagosomal proteins, the human Rab2 GTPase was detected interacting with a B. melitensis protein named RicA (Rab2 interacting conserved protein A). In silico analysis predict RicA protein to be 175 aa long, conserved in many bacteria group and possess features of a putative acetyltransferase. RicA was further characterized in this thesis.
Analyses in Y2H of randomly mutated alleles of ricA coding sequence indicate three regions are free of mutations, which is expected for the interaction interface. One of them contains a small segment that was called IGFP (Ile-Gly-Phe-Pro) loop. The IGFP loop was mutagenized at the Ile and Phe positions of this loop, and most of the mutants generated loss of interaction, suggesting that RicA interacts with Rab2 through the IGFP loop.
Rab2 3D structure analysis revealed two complementary hydrophobic patches which interestingly correspond to switch I and switch II regions, that are conformationally modified by GTP binding. This is consistent with apparently preferential binding of RicA to GDP bound form of GST-Rab2.Lors d’un crible de double hydride en levure (Y2H) à haut débit entre l’ORFéome de Brucella melitensis et des protéines potentiellement phagosomales humaines, la GTPase humaine Rab2 a été détectée comme interacteur d’une protéine de B. melitensis nommée RicA (Rab2 interacting conserved protein A). Une analyse in silico prédit que RicA est une protéine de 175 aa, conservée chez beaucoup de bactéries, et qui possède des caractéristiques d’acétyltransférases. RicA a été caractérisée au cours de ce travail de thèse.
Des analyses de Y2H avec des allèles aléatoirement mutés de la séquence codante de ricA ont indiqués que trois régions sont exemptes de mutations, ce qui est attendu pour une surface d’interaction. L’une d’elles contient un petit segment qui fut appellé boucle IGFP (Ile-Gly-Phe-Pro). La boucle IGFP fut mutée aux positions Ile et Phe de cette boucle, et la plupart des mutants ont généré des pertes d’interaction, suggérant que RicA interagit avec Rab2 grâce à cette boucle IGFP.
L’analyse de la structure tridimensionelle de Rab2 révèle deux surfaces hydrophobes complémentaires qui, de façon intéressante, correspondent aux régions “switch I” et “switch II”, dont la conformation est modifiée par la liaison de GTP. Ceci est cohérent avec une liaison apparemment préférentielle de RicA à la forme de GST-Rab2 liée au GDP.(DOCSC03) -- FUNDP, 200
Développement critique de méthodes d’analyse statistique de l’expression différentielle de gènes et de groupes de gènes, mesurée sur damiers à ADN
No full textDNA microarrays allow to study the expression profile of the whole genome of an organism. This technology is quite expensive, and the number of tested samples is often limited (at most 5 replicates). In those conditions, statistical tests are associated to bad performances. Various methods have been developed to optimize differential expression analysis. We describe a set of methods, to provide a comprehensive view of various approaches, of their main advantages and limitations.
Our first objective is to guide the analysis of differential expression, at the gene-level, by using biological or empirical informations. To improve the performances of statistical tests, we propose to share information across genes, gathered using an appropriate criteria. The window t-test has been developed following this strategy, to use the empirical relationship between variability and mean expression level. The window t-test only depends on the number of replicates, that defines the number of probesets used to compute variance estimates. Evaluation of methods reveals that the window t-test performs similarly to or better than the best methods [1].
Many biological informations can be used to define gene-sets (metabolic pathway, chromosomal location...). Current methods for gene-set analysis of differential expression are developed to test several hypothesis. We generalize gene-set analysis to answer the main Q0 question : « Does the individual expression values of the gene-set members differ between two condition ? ». We developed FAERI to answer to this question, by considering 3 criteria : the correlation between genes, the expression level, and the direction of the response (under- or over-expression). FAERI is a modified ANOVA-2 procedure, starting with a two-step reduction of expression data (Z-standardization, directional reduction). ANOVA-2 is shown to be the best-performing method when analyzing uni-directional gene-sets (all members are either activated, or repressed). FAERI reveals to be the most appropriate method for all tested gene-set types.
We developed PEGASE to perform differential expression analysis both at the gene and gene-set level. Consensus evaluation from several methods is included, to provide users with good results, even if the choice of an optimal method is not easy. Several methods are implemented in PEGASE, both at the gene and gene-set level, and performance evaluation can be run based on biological or empirical knowledge. PEGASE is also used as a back-end by PHOENIX, an online tool for microarray data analysis [2].
1. Berger F., De Hertogh B., Pierre M., Gaigneaux A. & Depiereux E. The "Window t-test": a simple and powerful approach to detect differentially expressed genes in microarray datasets. Cent. Eur. J. Biol., 2008, 3, 327-344.
2. Berger F., De Hertogh B., Bareke E., Pierre M., Gaigneaux A. & Depiereux E. PHOENIX: a web-interface for (re)analyses of microarray data. Cent. Eur. J. Biol., 2009, 4(4) : 603 : 618.Les puces à ADN permettent de mesurer le niveau d’expression de chacun des gènes du génome humain (ou d’autres organismes). Leur coût limite cependant le nombre de mesures réalisées, et les analyses statistiques traditionnelles sont peu performantes. Différentes méthodes ont été mises au point pour optimiser l'analyse de l'expression différentielle. Les plus représentatives sont décrites, pour faciliter la compréhension de la problématique, et des avantages et faiblesses présentées par les méthodes actuelles.
Notre première démarche vise à guider l'analyse sur base d'informations biologiques ou empiriques. Pour améliorer les performances de l'analyse, nous proposons de partager de l'information entre les gènes, groupés sur base d'un critère valide. La méthode window t-test a été mise au point en utilisant la relation empirique entre la variabilité et le niveau d'expression. Le window t-test dépends uniquement du nombre de mesures disponibles, qui détermine le nombre de gènes utilisés. Les performances obtenues avec un nombre limité de mesures (<=5) sont égales, voire supérieures, aux meilleurs méthodes actuelles [1].
Plusieurs critères biologiques permettent de définir des groupes de gènes (voie métabolique, localisation chromosomique...). Les méthodes actuelles d'analyse de groupes considèrent des hypothèses différentes. Notre démarche généralise l'étude des groupes par la question Q0 : « L'expression des membres du groupe est-elle différente entre deux conditions ? ». Nous avons conçu la méthode FAERI pour répondre à cette question en regard de trois critères : la corrélation des gènes, le niveau d'expression individuel, et la direction de la réponse (sur- ou sous-expression). FAERI utilise la procédure ANOVA-2, précédée par deux étapes de réduction des données (standardisation Z, réduction directionnelle). La méthode ANOVA-2 est la plus performante pour l'étude de groupes dont les membres sont activés et réprimés. La méthode FAERI est la plus appropriée pour tous les types de groupes testés.
Nos conclusions et notre démarche ont été automatisées, et matérialisées sous la forme du logiciel PEGASE. L'évaluation d'un consensus au départ des résultats de plusieurs méthodes d'analyse y est proposé, pour assurer à l'utilisateur l'obtention de résultats de bonne qualité, quelle que soit la méthode optimale. Plusieurs méthodes d'analyse sont proposées pour l'étude des gènes et des groupes, et les performances peuvent être évaluées sur base de données théoriques ou empiriques. PEGASE est utilisé par le serveur PHOENIX pour analyser les données [2].
1. Berger F., De Hertogh B., Pierre M., Gaigneaux A. & Depiereux E. The "Window t-test": a simple and powerful approach to detect differentially expressed genes in microarray datasets. Cent. Eur. J. Biol., 2008, 3, 327-344.
2. Berger F., De Hertogh B., Bareke E., Pierre M., Gaigneaux A. & Depiereux E. PHOENIX: a web-interface for (re)analyses of microarray data. Cent. Eur. J. Biol., 2009, 4(4) : 603 : 618.(DOCSC03) -- FUNDP, 200
Essays on Political Economy and Conflict Theory
No full textIn the first paper I study the incentives for electoral registration in a system
in which registration is costly. I argue that, if some more powerful agents in the
society can condition the voting behavior of part of the electorate, then more
easily controlled voters are also more likely to be registered. This allows the
powerful agents to have a large impact on election results, as the share of actual
votes controlled is increased through strategic registration. I show with the help
of a theoretical model that reducing the control on votes (for instance with the
adoption of a secret ballot) only partially reduces the bias in registration, as
scarcely motivated voters will be always easy to control. I test the predictions
of the model by examining in detail the effects of the introduction of the secret
ballot in Chile in 1958.
The second paper, coauthored with my colleague Petros Sekeris, investi-
gates the impact of land inequality on conflict intensity. We analyze how land
inequality across landlords influences the intensity of the fight against a rebel
group constituted by landless individuals. We show that conflict intensity is
non-monotonic in land inequality. In particular, the most severe conflicts occur
for intermediate land inequality levels. Moreover, under certain condition we
show that a Pareto improving transfer of land from the smaller to the larger
landlord exists.
Finally, the third paper also coauthored with Petros Sekeris, explores the
existence of deterrence equilibria in a general equilibrium model of “guns and
butter” production. In this class of model two agents choose how to allocate
their initial endowment between the production of consumables and weapons in
the first period of the game and whether to engage in war in the second period.
The standard result in the literature is that war is the only equilibrium of the
game. We show that, if fighting entails sufficiently low destruction, indeed war
is the unique equilibrium of the game. If, however, conflict generates sufficiently
large damages, agents start to adopt an alternative strategy: they arm to deter
their opponent from war. As a consequence, only mixed strategy equilibria
survive, in which players randomize over their deterrence and war strategies. In
this latter case, peace occurs with positive probability.(DOCSESG00 ) -- FUNDP, 200