Robust Nonlinear Mapping of Soil Contamination Using Support Regression

Spatial data analysis mapping and visualization is of great importance in various fields: environment, pollution, natural hazards and risks, epidemiology, spatial econometrics, etc. A basic task of spatial mapping is to make predictions based on some empirical data (measurements). A number of state-of-the-art methods can be used for the task: deterministic interpolations, methods of geostatistics: the family of kriging estimators (Deutsch and Journel, 1997), machine learning algorithms such as artificial neural networks (ANN) of different architectures, hybrid ANN-geostatistics models (Kanevski and Maignan, 2004; Kanevski et al., 1996), etc. All the methods mentioned above can be used for solving the problem of spatial data mapping. Environmental empirical data are always contaminated/corrupted by noise, and often with noise of unknown nature. That's one of the reasons why deterministic models can be inconsistent, since they treat the measurements as values of some unknown function that should be interpolated. Kriging estimators treat the measurements as the realization of some spatial randomn process. To obtain the estimation with kriging one has to model the spatial structure of the data: spatial correlation function or (semi-)variogram. This task can be complicated if there is not sufficient number of measurements and variogram is sensitive to outliers and extremes. ANN is a powerful tool, but it also suffers from the number of reasons. of a special type ? multiplayer perceptrons ? are often used as a detrending tool in hybrid (ANN+geostatistics) models (Kanevski and Maignank, 2004). Therefore, development and adaptation of the method that would be nonlinear and robust to noise in measurements, would deal with the small empirical datasets and which has solid mathematical background is of great importance. The present paper deals with such model, based on Statistical Learning Theory (SLT) - Support Vector Regression. SLT is a general mathematical framework devoted to the problem of estimation of the dependencies from empirical data (Hastie et al, 2004; Vapnik, 1998). SLT models for classification - Support Vector Machines - have shown good results on different machine learning tasks. The results of SVM classification of spatial data are also promising (Kanevski et al, 2002). The properties of SVM for regression - Support Vector Regression (SVR) are less studied. First results of the application of SVR for spatial mapping of physical quantities were obtained by the authorsin for mapping of medium porosity (Kanevski et al, 1999), and for mapping of radioactively contaminated territories (Kanevski and Canu, 2000). The present paper is devoted to further understanding of the properties of SVR model for spatial data analysis and mapping. Detailed description of the SVR theory can be found in (Cristianini and Shawe-Taylor, 2000; Smola, 1996) and basic equations for the nonlinear modeling are given in section 2. Section 3 discusses the application of SVR for spatial data mapping on the real case study - soil pollution by Cs137 radionuclide. Section 4 discusses the properties of the modelapplied to noised data or data with outliers.

Etalonnage de trois épreuves de recognition continue : Effets de l'âge, du sexe et du niveau socio-professionnel sur les mesures de discrimination et de biais de réponse. [Standardization of three continuous recognition test forms - Effects of age, sex and socio-economic level on measurements of discrimination and response bias.]

Nous présontons l'étalonnage d'un test mnésique de recognition dans un échantillon de 180 adultes francophones de la Suisse Romande. Le test comprend trois formes utilisant un matériel verbal (mots) ou non verbal (visages ou paysages). Une attention particulière est accordée à l'âge dans la présentation des résultats. Celui-ci affecte plus précocement et plus intensément la performance aux formes non verbales qu'à la forme verbale du test. Il induit également une importante augmentation du nombre de fausses reconnaissances pour les formes non verbales.

Characterising the role of T-cell subsets in experimental allograft rejection and transplantation tolerance

SUMMARY : The recognition by recipient T cells of the allograft major histocompatibility complex (MHC)mismatched antigens is the primary event that ultimately leads to rejection. In the transplantation setting, circulating alloreactive CD4+ T cells play a central role in the initiation and the coordination of the immune response and can initiate the rejection of an allograft via three distinct pathways: the direct, indirect and the recently described semi-direct pathway. However, the exact role of individual CD4+ T-cell subsets in the development of allograft rejection is not clearly defined. Furthermore, besides pathogenic effector T cells, a new subset of T cells with regulatory properties, the CD4+CD25+Foxp3+ (Treg) cells, has come under increased scrutiny over the last decade. The experiments presented in this thesis were designed to better define the phenotype and functional characteristics of CD4+ T-cell subsets and Treg cells in vitro and in vivo in a marine adoptive transfer and skin transplantation model. As Treg cells play a key role in the induction and maintenance of peripheral transplantation tolerance, we have explored whether donor-antigen specific Treg cells could be expanded in vitro. Here we describe a robust protocol for the ex-vivo generation and expansion of antigen-specific Treg cells, without loss of their characteristic phenotype and suppressive function. In our in vivo transplantation model, antigen-specific Treg cells induced donor-specific tolerance to skin allografts in lymphopenic recipients and significantly delayed skin graft rejection in wild-type mice in the absence of any other immunosuppression. Naïve and memory CD4+ T cells have distinct phenotypes, effector functions and in vivo homeostatsis, and thus may play different roles in anti-donor immunity after transplantation. We have analyzed in vitro and in vivo primary alloresponses of naïve and cross-reactive memory CD4+ T cells. We found that the CD4+CD45RBlo memory T-cell pool was heterogeneous and contained cells with regulatory potentials, both in the CD4+CD25+ and CD4+CD25- populations. CD4+ T cells capable of inducing strong primary alloreactive responses in vitro and rejection of a first allograft in vivo were mainly contained within the CD45RBhi naïve CD4+ T-cell compartment. Taken together, the work described in this thesis provides new insights into the mechanisms that drive allograft rejection or donor-specific transplantation tolerance. These results will help to optimise current clinical immunosuppressive regimens used after solid organ transplantation and design new immunotherapeutic strategies to prevent transplant rejection. RÉSUMÉ : ROLE DES SOUS-POPULATIONS DE CELLULES T DANS LE REJET DE GREFFE ET L'INDUCTION DE TOLERANCE EN TRANSPLANTATION La reconnaissance par les cellules T du receveur des alloantigènes du complexe majeur d'histocompatibilité (CMIT) présentés par une greffe allogénique, est le premier événement qui aboutira au rejet de l'organe greffé. Dans le contexte d'une transplantation, les cellules alloréactives T CD4+ circulantes jouent un rôle central dans l'initiation et la coordination de 1a réponse immune, et peuvent initier le rejet par 3 voies distinctes : la voie directe, indirecte et la voie servi-directe, plus récemment décrite. Toutefois, le rôle exact des sous-populations de cellules T CD4+ dans les différentes étapes menant au rejet d'une allogreffe n'est pas clairement établi. Par ailleurs, hormis les cellules T effectrices pathogéniques, une sous-population de cellules T ayant des propriétés régulatrices, les cellules T CD4+CD25+Foxp3+ (Treg), a été nouvellement décrite et est intensément étudiée depuis environ dix ans. Les expériences présentées dans cette thèse ont été planifiées afin de mieux définir le phénotype et les caractéristiques fonctionnels des sous-populations de cellules T CD4+ et des Treg in vitro et in vivo dans un modèle marin de transfert adoptif de cellules et de transplantation de peau. Comme les cellules Treg jouent un rôle clé dans l'induction et le maintien de la tolérance périphérique en transplantation, nous avons investigué la possibilité de multiplier in vitro des cellules Treg avec spécificité antigénique pour le donneur. Nous décrivons ici un protocole reproductible pour la génération et l'expansion ex-vivo de cellules Treg avec spécificité antigénique, sans perte de leur phénotype caractéristique et de leur fonction suppressive. Dans notre modèle in vivo de transplantation de peau, ces cellules Treg pouvaient induire une tolérance spécifique vis-à-vis du donneur chez des souris lymphopéniques, et, chez des souris normales non-lymphopéniques ces Treg ont permis de retarder significativement le rejet en l'absence de tout traitement immunosuppresseur. Les cellules T CD4+ naïves et mémoires se distinguent par leur phénotype, fonction effectrice et leur homéostasie in vivo, et peuvent donc moduler différemment la réponse immune contre le donneur après transplantation. Nous avons analysé in vitro et in vivo les réponses allogéniques primaires de cellules T CD4+ naïves et mémoires non-spécifiques (cross-réactives). Nos résultats ont montré que le pool de cellules T CD4+CD45RB'° mémoires était hétérogène et contenait des cellules avec un potentiel régulateur, aussi bien parmi la sous-population de cellules CD4+CD25+ que CD4+CD25+. Les cellules T CD4+ capables d'induire une alloréponse primaire intense in vitro et le rejet d'une première allogreffe in vivo étaient essentiellement contenues dans le pool de cellules T CD4+CD45RBhi naïves. En conclusion, le travail décrit dans cette thèse amène un nouvel éclairage sur les mécanismes responsables du rejet d'une allogreffe ou de l'induction de tolérance en transplantation. Ces résultats permettront d'optimaliser les traitements immunosuppresseurs utilisés en transplantation clinique et de concevoir des nouvelles stratégies irnmuno-thérapeutiques pour prévenir le rejet de greffe allogénique.

Comparison and validation of tissue modelization and statistical classification methods in T1-weighted MR brain images.

This paper presents a validation study on statistical nonsupervised brain tissue classification techniques in magnetic resonance (MR) images. Several image models assuming different hypotheses regarding the intensity distribution model, the spatial model and the number of classes are assessed. The methods are tested on simulated data for which the classification ground truth is known. Different noise and intensity nonuniformities are added to simulate real imaging conditions. No enhancement of the image quality is considered either before or during the classification process. This way, the accuracy of the methods and their robustness against image artifacts are tested. Classification is also performed on real data where a quantitative validation compares the methods' results with an estimated ground truth from manual segmentations by experts. Validity of the various classification methods in the labeling of the image as well as in the tissue volume is estimated with different local and global measures. Results demonstrate that methods relying on both intensity and spatial information are more robust to noise and field inhomogeneities. We also demonstrate that partial volume is not perfectly modeled, even though methods that account for mixture classes outperform methods that only consider pure Gaussian classes. Finally, we show that simulated data results can also be extended to real data.

Unicoloniality, recognition and genetic differentiation in a native Formica ant.

Some ants have an extraordinary form of social organization, called unicoloniality, whereby individuals mix freely among physically separated nests. This mode of social organization has been primarily studied in introduced and invasive ant species, so that the recognition ability and genetic structure of ants forming unicolonial populations in their native range remain poorly known. We investigated the pattern of aggression and the genetic structure of six unicolonial populations of the ant Formica paralugubris at four hierarchical levels: within nests, among nests within the same population, among nests of populations within the Alps or Jura Mountains and among nests of the two mountain ranges. Ants within populations showed no aggressive behaviour, but recognized nonnestmates as shown by longer antennation bouts. Overall, the level of aggression increased with geographic and genetic distance but was always considerably lower than between species. No distinct behavioural supercolony boundaries were found. Our study provides evidence that unicoloniality can be maintained in noninvasive ants despite significant genetic differentiation and the ability to discriminate between nestmates and nonnestmates.

On the evolution of harming and recognition in finite panmictic and infinite structured populations.

Natural selection may favor two very different types of social behaviors that have costs in vital rates (fecundity and/or survival) to the actor: helping behaviors, which increase the vital rates of recipients, and harming behaviors, which reduce the vital rates of recipients. Although social evolutionary theory has mainly dealt with helping behaviors, competition for limited resources creates ecological conditions in which an actor may benefit from expressing behaviors that reduce the vital rates of neighbors. This may occur if the reduction in vital rates decreases the intensity of competition experienced by the actor or that experienced by its offspring. Here, we explore the joint evolution of neutral recognition markers and marker-based costly conditional harming whereby actors express harming, conditional on actor and recipient bearing different conspicuous markers. We do so for two complementary demographic scenarios: finite panmictic and infinite structured populations. We find that marker-based conditional harming can evolve under a large range of recombination rates and group sizes under both finite panmictic and infinite structured populations. A direct comparison with results for the evolution of marker-based conditional helping reveals that, if everything else is equal, marker-based conditional harming is often more likely to evolve than marker-based conditional helping.

Reconciling phonological neighborhood effects in speech production through single trial analysis

A crucial step for understanding how lexical knowledge is represented is to describe the relative similarity of lexical items, and how it influences language processing. Previous studies of the effects of form similarity on word production have reported conflicting results, notably within and across languages. The aim of the present study was to clarify this empirical issue to provide specific constraints for theoretical models of language production. We investigated the role of phonological neighborhood density in a large-scale picture naming experiment using fine-grained statistical models. The results showed that increasing phonological neighborhood density has a detrimental effect on naming latencies, and re-analyses of independently obtained data sets provide supplementary evidence for this effect. Finally, we reviewed a large body of evidence concerning phonological neighborhood density effects in word production, and discussed the occurrence of facilitatory and inhibitory effects in accuracy measures. The overall pattern shows that phonological neighborhood generates two opposite forces, one facilitatory and one inhibitory. In cases where speech production is disrupted (e.g. certain aphasic symptoms), the facilitatory component may emerge, but inhibitory processes dominate in efficient naming by healthy speakers. These findings are difficult to accommodate in terms of monitoring processes, but can be explained within interactive activation accounts combining phonological facilitation and lexical competition.

Nuclear translocation and DNA recognition signals colocalized within the bZIP domain of cyclic adenosine 3',5'-monophosphate response element-binding protein CREB.

CREB is a cAMP-responsive nuclear DNA-binding protein that binds to cAMP response elements and stimulates gene transcription upon activation of the cAMP signalling pathway. The protein consists of an amino-terminal transcriptional transactivation domain and a carboxyl-terminal DNA-binding domain (bZIP domain) comprised of a basic region and a leucine zipper involved in DNA recognition and dimerization, respectively. Recently, we discovered a testis-specific transcript of CREB that contains an alternatively spliced exon encoding multiple stop codons. CREB encoded by this transcript is a truncated protein lacking the bZIP domain. We postulated that the antigen detected by CREB antiserum in the cytoplasm of germinal cells is the truncated CREB that must also lack its nuclear translocation signal (NTS). To test this hypothesis we prepared multiple expression plasmids encoding carboxyl-terminal deletions of CREB and transiently expressed them in COS-1 cells. By Western immunoblot analysis as well as immunocytochemistry of transfected cells, we show that CREB proteins truncated to amino acid 286 or shorter are sequestered in the cytoplasm, whereas a CREB of 295 amino acids is translocated into the nucleus. Chimeric CREBs containing a heterologous NTS fused to the first 248 or 261 amino acids of CREB are able to drive the translocation of the protein into the nucleus. Thus, the nine amino acids in the basic region involved in DNA recognition between positions 287 and 295 (RRKKKEYVK) of CREB contain the NTS. Further, mutation of the lysine at position 290 in CREB to an asparagine diminishes nuclear translocation of the protein.(ABSTRACT TRUNCATED AT 250 WORDS)

Membrane insertion and antibody recognition of a glycosylphosphatidylinositol-anchored protein: an optical study.

The kinetics of binding of a glycolipid-anchored protein (the promastigote surface protease, PSP) to planar lecithin bilayers is studied by an integrated optics technique, in which the bilayer membrane is supported on an optical wave guide and the phase velocities of guided light modes in the wave guide are measured. From these velocities, the optical parameters of the membrane and PSP layers deposited on the waveguide are determined, yielding in particular the mass of PSP bound to the membrane, which is followed in real time. From a comparison of the binding rates of PSP and PSP from which the lipid moiety has been removed, it is shown that the lipid moiety plays a key role in anchoring the protein to the membrane. Specific and nonspecific binding of antibodies to membrane-anchored PSP is also investigated. As little as a fifth of a monolayer of PSP is sufficient to suppress the appreciable nonspecific binding of antibodies to the membrane.

Recognition of Gram-positive Intestinal Bacteria by Hybridoma- and Colostrum-derived Secretory Immunoglobulin A Is Mediated by Carbohydrates.

Humans live in symbiosis with 10(14) commensal bacteria among which >99% resides in their gastrointestinal tract. The molecular bases pertaining to the interaction between mucosal secretory IgA (SIgA) and bacteria residing in the intestine are not known. Previous studies have demonstrated that commensals are naturally coated by SIgA in the gut lumen. Thus, understanding how natural SIgA interacts with commensal bacteria can provide new clues on its multiple functions at mucosal surfaces. Using fluorescently labeled, nonspecific SIgA or secretory component (SC), we visualized by confocal microscopy the interaction with various commensal bacteria, including Lactobacillus, Bifidobacteria, Escherichia coli, and Bacteroides strains. These experiments revealed that the interaction between SIgA and commensal bacteria involves Fab- and Fc-independent structural motifs, featuring SC as a crucial partner. Removal of glycans present on free SC or bound in SIgA resulted in a drastic drop in the interaction with Gram-positive bacteria, indicating the essential role of carbohydrates in the process. In contrast, poor binding of Gram-positive bacteria by control IgG was observed. The interaction with Gram-negative bacteria was preserved whatever the molecular form of protein partner used, suggesting the involvement of different binding motifs. Purified SIgA and SC from either mouse hybridoma cells or human colostrum exhibited identical patterns of recognition for Gram-positive bacteria, emphasizing conserved plasticity between species. Thus, sugar-mediated binding of commensals by SIgA highlights the currently underappreciated role of glycans in mediating the interaction between a highly diverse microbiota and the mucosal immune system.

T cell recognition of hapten. Anatomy of T cell receptor binding of a H-2kd-associated photoreactive peptide derivative.

To elucidate the structural basis of T cell recognition of hapten-modified antigenic peptides, we studied the interaction of the T1 T cell antigen receptor (TCR) with its ligand, the H-2Kd-bound Plasmodium berghei circumsporozoite peptide 252-260 (SYIPSAEKI) containing photoreactive 4-azidobenzoic acid (ABA) on P. berghei circumsporozoite Lys259. The photoaffinity-labeled TCR residue(s) were mapped as Tyr48 and/or Tyr50 of complementary determining region 2beta (CDR2beta). Other TCR-ligand contacts were identified by mutational analysis. Molecular modeling, based on crystallographic coordinates of closely related TCR and major histocompatibility complex I molecules, indicated that ABA binds strongly and specifically in a cavity between CDR3alpha and CDR2beta. We conclude that TCR expressing selective Vbeta and CDR3alpha sequences form a binding domain between CDR3alpha and CDR2beta that can accommodate nonpeptidic moieties conjugated at the C-terminal portion of peptides binding to major histocompatibility complex (MHC) encoded proteins.

Robust Detection of Incumbents in Cognitive Radio Networks Using Groups

Cognitive radio is a wireless technology aimed at improvingthe efficiency use of the radio-electric spectrum, thus facilitating a reductionin the load on the free frequency bands. Cognitive radio networkscan scan the spectrum and adapt their parameters to operate in the unoccupiedbands. To avoid interfering with licensed users operating on a givenchannel, the networks need to be highly sensitive, which is achieved byusing cooperative sensing methods. Current cooperative sensing methodsare not robust enough against occasional or continuous attacks. This articleoutlines a Group Fusion method that takes into account the behavior ofusers over the short and long term. On fusing the data, the method is basedon giving more weight to user groups that are more unanimous in their decisions.Simulations have been performed in a dynamic environment withinterferences. Results prove that when attackers are present (both reiterativeor sporadic), the proposed Group Fusion method has superior sensingcapability than other methods.