Mixture of bivariate Poisson regression models with an application to insurance

In a recent paper Bermúdez [2009] used bivariate Poisson regression models for ratemaking in car insurance, and included zero-inflated models to account for the excess of zeros and the overdispersion in the data set. In the present paper, we revisit this model in order to consider alternatives. We propose a 2-finite mixture of bivariate Poisson regression models to demonstrate that the overdispersion in the data requires more structure if it is to be taken into account, and that a simple zero-inflated bivariate Poisson model does not suffice. At the same time, we show that a finite mixture of bivariate Poisson regression models embraces zero-inflated bivariate Poisson regression models as a special case. Additionally, we describe a model in which the mixing proportions are dependent on covariates when modelling the way in which each individual belongs to a separate cluster. Finally, an EM algorithm is provided in order to ensure the models’ ease-of-fit. These models are applied to the same automobile insurance claims data set as used in Bermúdez [2009] and it is shown that the modelling of the data set can be improved considerably.

On the stability of the optimal value and the optimal set in optimization problems

The paper develops a stability theory for the optimal value and the optimal set mapping of optimization problems posed in a Banach space. The problems considered in this paper have an arbitrary number of inequality constraints involving lower semicontinuous (not necessarily convex) functions and one closed abstract constraint set. The considered perturbations lead to problems of the same type as the nominal one (with the same space of variables and the same number of constraints), where the abstract constraint set can also be perturbed. The spaces of functions involved in the problems (objective and constraints) are equipped with the metric of the uniform convergence on the bounded sets, meanwhile in the space of closed sets we consider, coherently, the Attouch-Wets topology. The paper examines, in a unified way, the lower and upper semicontinuity of the optimal value function, and the closedness, lower and upper semicontinuity (in the sense of Berge) of the optimal set mapping. This paper can be seen as a second part of the stability theory presented in [17], where we studied the stability of the feasible set mapping (completed here with the analysis of the Lipschitz-like property).

Persistence of EBV Antigen-Specific CD8 T Cell Clonotypes during Homeostatic Immune Reconstitution in Cancer Patients.

Persistent viruses are kept in check by specific lymphocytes. The clonal T cell receptor (TCR) repertoire against Epstein-Barr virus (EBV), once established following primary infection, exhibits a robust stability over time. However, the determinants contributing to this long-term persistence are still poorly characterized. Taking advantage of an in vivo clinical setting where lymphocyte homeostasis was transiently perturbed, we studied EBV antigen-specific CD8 T cells before and after non-myeloablative lympho-depleting chemotherapy of melanoma patients. Despite more advanced T cell differentiation, patients T cells showed clonal composition comparable to healthy individuals, sharing a preference for TRBV20 and TRBV29 gene segment usage and several co-dominant public TCR clonotypes. Moreover, our data revealed the presence of relatively few dominant EBV antigen-specific T cell clonotypes, which mostly persisted following transient lympho-depletion (TLD) and lymphocyte recovery, likely related to absence of EBV reactivation and de novo T cell priming in these patients. Interestingly, persisting clonotypes frequently co-expressed memory/homing-associated genes (CD27, IL7R, EOMES, CD62L/SELL and CCR5) supporting the notion that they are particularly important for long-lasting CD8 T cell responses. Nevertheless, the clonal composition of EBV-specific CD8 T cells was preserved over time with the presence of the same dominant clonotypes after non-myeloablative chemotherapy. The observed clonotype persistence demonstrates high robustness of CD8 T cell homeostasis and reconstitution.

Suppression of Arabidopsis protophloem differentiation and root meristem growth by CLE45 requires the receptor-like kinase BAM3.

Peptide signaling presumably occupies a central role in plant development, yet only few concrete examples of receptor-ligand pairs that act in the context of specific differentiation processes have been described. Here we report that second-site null mutations in the Arabidopsis leucine-rich repeat receptor-like kinase gene barely any meristem 3 (BAM3) perfectly suppress the postembryonic root meristem growth defect and the associated perturbed protophloem development of the brevis radix (brx) mutant. The roots of bam3 mutants specifically resist growth inhibition by the CLAVATA3/ENDOSPERM SURROUNDING REGION 45 (CLE45) peptide ligand. WT plants transformed with a construct for ectopic overexpression of CLE45 could not be recovered, with the exception of a single severely dwarfed and sterile plant that eventually died. By contrast, we obtained numerous transgenic bam3 mutants transformed with the same construct. These transgenic plants displayed a WT phenotype, however, supporting the notion that CLE45 is the likely BAM3 ligand. The results correlate with the observation that external CLE45 application represses protophloem differentiation in WT, but not in bam3 mutants. BAM3, BRX, and CLE45 are expressed in a similar spatiotemporal trend along the developing protophloem, up to the end of the transition zone. Induction of BAM3 expression upon CLE45 application, ectopic overexpression of BAM3 in brx root meristems, and laser ablation experiments suggest that intertwined regulatory activity of BRX, BAM3, and CLE45 could be involved in the proper transition of protophloem cells from proliferation to differentiation, thereby impinging on postembryonic growth capacity of the root meristem.

Statistical treatment of grain-size curves and empirical distributions: densities as compositions?

The preceding two editions of CoDaWork included talks on the possible considerationof densities as infinite compositions: Egozcue and D´ıaz-Barrero (2003) extended theEuclidean structure of the simplex to a Hilbert space structure of the set of densitieswithin a bounded interval, and van den Boogaart (2005) generalized this to the setof densities bounded by an arbitrary reference density. From the many variations ofthe Hilbert structures available, we work with three cases. For bounded variables, abasis derived from Legendre polynomials is used. For variables with a lower bound, westandardize them with respect to an exponential distribution and express their densitiesas coordinates in a basis derived from Laguerre polynomials. Finally, for unboundedvariables, a normal distribution is used as reference, and coordinates are obtained withrespect to a Hermite-polynomials-based basis.To get the coordinates, several approaches can be considered. A numerical accuracyproblem occurs if one estimates the coordinates directly by using discretized scalarproducts. Thus we propose to use a weighted linear regression approach, where all k-order polynomials are used as predictand variables and weights are proportional to thereference density. Finally, for the case of 2-order Hermite polinomials (normal reference)and 1-order Laguerre polinomials (exponential), one can also derive the coordinatesfrom their relationships to the classical mean and variance.Apart of these theoretical issues, this contribution focuses on the application of thistheory to two main problems in sedimentary geology: the comparison of several grainsize distributions, and the comparison among different rocks of the empirical distribution of a property measured on a batch of individual grains from the same rock orsediment, like their composition

Modelling structural zeros in compositional data

This analysis was stimulated by the real data analysis problem of householdexpenditure data. The full dataset contains expenditure data for a sample of 1224 households. The expenditure is broken down at 2 hierarchical levels: 9 major levels (e.g. housing, food, utilities etc.) and 92 minor levels. There are also 5 factors and 5 covariates at the household level. Not surprisingly, there are a small number of zeros at the major level, but many zeros at the minor level. The question is how best to model the zeros. Clearly, models that tryto add a small amount to the zero terms are not appropriate in general as at least some of the zeros are clearly structural, e.g. alcohol/tobacco for households that are teetotal. The key question then is how to build suitable conditional models. For example, is the sub-composition of spendingexcluding alcohol/tobacco similar for teetotal and non-teetotal households?In other words, we are looking for sub-compositional independence. Also, what determines whether a household is teetotal? Can we assume that it is independent of the composition? In general, whether teetotal will clearly depend on the household level variables, so we need to be able to model this dependence. The other tricky question is that with zeros on more than onecomponent, we need to be able to model dependence and independence of zeros on the different components. Lastly, while some zeros are structural, others may not be, for example, for expenditure on durables, it may be chance as to whether a particular household spends money on durableswithin the sample period. This would clearly be distinguishable if we had longitudinal data, but may still be distinguishable by looking at the distribution, on the assumption that random zeros will usually be for situations where any non-zero expenditure is not small.While this analysis is based on around economic data, the ideas carry over tomany other situations, including geological data, where minerals may be missing for structural reasons (similar to alcohol), or missing because they occur only in random regions which may be missed in a sample (similar to the durables)

ELECTROCORTICAL DYNAMICS OF MOTOR INHIBITORY CONTROL WITH AGING

Motor inhibitory control plays a central role in adaptive behaviors during the entire lifespan. Inhibitory motor control refers to the ability to stop all (global) or a part (selective) of a planned or ongoing motor action. Although the neural processing underlying the global inhibitory control has received much attention from cognitive neuroscientists, brain modulations that occur during selective inhibitory motor control remain unknown. The aim of the present thesis is to investigate the spatio-temporal brain processes of selective inhibitory motor control in young and old adults using high-density electroencephalography. In the first part, we focus on early (preparatory period) spatio-temporal brain processes involved in selective and global inhibitory control in young (study I) and old adults (study II) using a modified Go/No-go task. In study I, we distinguished global from selective inhibition in the early attentional stage of inhibitory control and provided neurophysiological evidence in favor of the combination model. In study II, we showed an under-recruitment of neural resources associated with preservation of performance in old adults during selective inhibition, suggesting efficient cerebral and behavioral adaptations to environmental changes. In the second part, we investigate beta oscillations in the late (post-execution period) spatio-temporal brain processes of selective inhibition during a motor Switching task (i.e., tapping movement from bimanual to unimanual) in young (study III) and old adults (study IV). In study III, we identified concomitant beta synchronization related (i) to sensory reafference processes, which enabled the stabilization of the movement that was perturbed after switching, and (ii) to active inhibition processes that prevented movement of the stopping hand. In study IV, we demonstrated a larger beta synchronization in frontal and parietal regions in old adults compared to young adults, suggesting age-related brain modulations in active inhibition processes. Apart from contributing to a basic understanding of the electrocortical dynamics underlying inhibitory motor control, the findings of the present studies contribute to knowledge regarding the further establishment of specific trainings with aging. -- Le contrôle de l'inhibition motrice joue un rôle central dans les adaptations comportementales quel que soit l'âge. L'inhibition motrice se réfère à la capacité à arrêter entièrement (globale) ou en partie (sélective) une action motrice planifiée ou en cours. Bien que les processus neuronaux sous-jacents de l'inhibition globale aient suscité un grand intérêt auprès des neurosciences cognitives, les modulations cérébrales dans le contrôle de l'inhibition motrice sélective sont encore peu connues. Le but de cette thèse est d'étudier les processus cérébraux spatio-temporels du contrôle de l'inhibition motrice sélective chez les adultes jeunes et âgés en utilisant l'électroencéphalogramme à haute densité. Dans la première partie, nous comparons les processus cérébraux spatio-temporels précoces (préparation motrice) de l'inhibition sélective et globale chez des adultes jeunes (étude I) et âgés (étude II) en utilisant une tâche Go/No-go modifiée. Dans l'étude I, nous avons distingué l'inhibition globale et sélective au niveau des processus attentionnels précoces du contrôle de l'inhibition et nous avons apporté des preuves neurophysiologiques de l'existence d'un modèle de combinaison. Dans l'étude II, nous avons montré une sous-activation neuronale associée à un maintien de la performance dans l'inhibition sélective chez les adultes âgés, suggérant des adaptations cérébrales et comportementales aux contraintes environnementales. Dans la seconde partie, nous examinons les processus cérébraux spatio-temporels tardifs (post-exécution motrice) de l'inhibition sélective pendant une tâche de Switching (tapping bimanuel vers un tapping unimanuel) chez des adultes jeunes (étude III) et âgés (étude IV). Dans l'étude III, nous avons distingué des synchronisations beta liées (i) au traitement des réafférences sensorielles permettant de stabiliser le mouvement perturbé après le switching, et (ii) aux processus d'inhibition active afin d'empêcher les mouvements de la main arrêtée. Dans l'étude IV, cette synchronisation beta était plus forte dans les régions frontales et pariétales chez les âgés par rapport aux jeunes adultes suggérant des modulations cérébrales de l'inhibition active avec l'âge. Outre la contribution fondamentale sur la compréhension des dynamiques électrocorticales dans le contrôle de l'inhibition motrice, les résultats de ces études contribuent à développer les connaissances pour la mise en place de programmes d'entraînements adaptés aux personnes âgées.

A compositional statistical analysis of capital stock

Most of economic literature has presented its analysis under the assumption of homogeneous capital stock.However, capital composition differs across countries. What has been the pattern of capital compositionassociated with World economies? We make an exploratory statistical analysis based on compositional datatransformed by Aitchinson logratio transformations and we use tools for visualizing and measuring statisticalestimators of association among the components. The goal is to detect distinctive patterns in the composition.As initial findings could be cited that:1. Sectorial components behaved in a correlated way, building industries on one side and , in a lessclear view, equipment industries on the other.2. Full sample estimation shows a negative correlation between durable goods component andother buildings component and between transportation and building industries components.3. Countries with zeros in some components are mainly low income countries at the bottom of theincome category and behaved in a extreme way distorting main results observed in the fullsample.4. After removing these extreme cases, conclusions seem not very sensitive to the presence ofanother isolated cases

Possible solution of some essential zero problems in compositional data analysis

One of the tantalising remaining problems in compositional data analysis lies in how to deal with data sets in which there are components which are essential zeros. By anessential zero we mean a component which is truly zero, not something recorded as zero simply because the experimental design or the measuring instrument has not been sufficiently sensitive to detect a trace of the part. Such essential zeros occur inmany compositional situations, such as household budget patterns, time budgets,palaeontological zonation studies, ecological abundance studies. Devices such as nonzero replacement and amalgamation are almost invariably ad hoc and unsuccessful insuch situations. From consideration of such examples it seems sensible to build up amodel in two stages, the first determining where the zeros will occur and the secondhow the unit available is distributed among the non-zero parts. In this paper we suggest two such models, an independent binomial conditional logistic normal model and a hierarchical dependent binomial conditional logistic normal model. The compositional data in such modelling consist of an incidence matrix and a conditional compositional matrix. Interesting statistical problems arise, such as the question of estimability of parameters, the nature of the computational process for the estimation of both the incidence and compositional parameters caused by the complexity of the subcompositional structure, the formation of meaningful hypotheses, and the devising of suitable testing methodology within a lattice of such essential zero-compositional hypotheses. The methodology is illustrated by application to both simulated and real compositional data

Speed of reaction-diffusion fronts in spatially heterogeneous media

The front speed problem for nonuniform reaction rate and diffusion coefficient is studied by using singular perturbation analysis, the geometric approach of Hamilton-Jacobi dynamics, and the local speed approach. Exact and perturbed expressions for the front speed are obtained in the limit of large times. For linear and fractal heterogeneities, the analytic results have been compared with numerical results exhibiting a good agreement. Finally we reach a general expression for the speed of the front in the case of smooth and weak heterogeneities

Molecular role of Cx37 in advanced atherosclerosis: a micro-array study.

Recently, we showed that connexin37 (Cx37) protects against early atherosclerotic lesion development by regulating monocyte adhesion. The expression of this gap junction protein is altered in mouse and human atherosclerotic lesions; it is increased in macrophages newly recruited to the lesions and disappears from the endothelium of advanced plaques. To obtain more insight into the molecular role of Cx37 in advanced atherosclerosis, we used micro-array analysis for gene expression profiling in aortas of ApoE(-/-) and Cx37(-/-)ApoE(-/-) mice before and after 18 weeks of cholesterol-rich diet. Out of >15,000 genes, 106 genes were significantly differentially expressed in young mice before diet (P-value of <0.05, fold change of >0.7 or <-0.7, and intensity value >2.2 times background). Ingenuity pathway analysis (IPA) revealed differences in genes involved in cell-to-cell signaling and interaction, cellular compromise and nutritional disease. In addition, we identified 100 genes that were significantly perturbed after the cholesterol-rich diet. Similar to the analysis on 10-week-old mice, IPA revealed differences in genes involved in cell-to-cell signaling and interaction as well as to immuno-inflammatory disease. Furthermore, we found important changes in genes involved in vascular calcification and matrix degradation, some of which were confirmed at protein level by (immuno-)histochemistry. In conclusion, we suggest that Cx37 deficiency alters the global differential gene expression profiles in young mice towards a pro-inflammatory phenotype, which are then further influenced in advanced atherosclerosis. The results provide new insights into the significance of Cx37 in plaque calcification.

Characterization of N-cadherin association to lipid rafts in melanoma

RESUME La peau est un organe complex composé de deux parties distinctes: l'épiderme et le derme, séparé par une membrane basale. Dans la couche basale de l'épiderme, les melanocytes synthétisent la mélanine dans des mélanosomes. Les mélanosomes sont ensuite transportés des mélanocytes vers les kératinocytes, protégeant ainsi la peau des dégâts dus aux radiations U.V. La E-cadhérine assure l'adhésion entre les mélanocytes et les kératinocytes. Au cours de la transformation du mélanocyte en cellule malignes, les mélanocytes perdent l'expression de la E-cadhérine et, simultanément, se mettent à exprimer la N-cadhérine, ce phénomène est nommé « cadherin switch ». La perte de l'expression de la E-cadhérine permet au mélanocytes d'échapper au contrôle des kératinocytes, tandis que l'expression de la N-cadhérine promeut l'invasion métastasique des cellules de mélanome. Préalablement, nous avons trouvé qu'une fraction de la N-cadhérine était localisée les microdomaines membranaires spécialisés, enrichi en cholestérol et en glycosphingolipides, appelés « lipid rafts ». Une des particularité des « lipid rafts » est qu'ils sont riches en molécules permettant la transmission de signaux d'activation. De plus, des travaux récents rapportent qu'un sous-type de « lipid rafts » appelé caveolae pourrai contribuer à la progression tumorale. S'appuyant sur le rôle prépondérant de la N-cadhérine dans la progression du mélanome ainsi que sur sa présence dans les « lipid rafts », nous avons émis l'hypothèse que l'association de la N-cadhérine avec les « lipid rafts » pourrai contribuer à la progression du mélanome. Le but de ce projet à été de caractériser l'association de la Ncadhérine avec les « lipid rafts » au cours de la progression du mélanome. Au moyen de lignées cellulaires humaines, dérivées de mélanomes à différents stades de progression, nous avons trouvé que (1) la N-cadhérine est partiellement associée aux «lipid rafts » dans six lignées dérivées de mélanome en phase avancée de progression et dans des tumeurs expérimentales, mais pas dans deux lignées dérivées de mélanome à un stade plus précoce ; (2) l'association de la N-cadhérine dans les « lipid rafts » ne dépent pas de son niveau d'expression ; (3) la E-cadhérine n'est pas présente dans les « lipid rafts »d'une lignée de cellule de mélanome ayant conservé l'expression de la E-cadhérine ; (4) la localisation de la N-cadhérine dans les « lipid rafts »n'est pas modulée par les facteurs de croissance bFGF, IGF-I, et HRG1-β1, ni par des voies de signalisation impliquant MEK, PKA, les kinases de la famille Src, et PI3K ; (5) l'association de la N-cadhérine avec les « lipid rafts » n'est pas requise pour la stabilisation des jonctions adhérentes et n'est pas perturbée par la destruction de ces dernières ; (6) la N-cadhérine dans les « lipid rafts » forme un complexe avec β-caténine, p 120ctn et α-caténine. En conclusion, cette étude originale montre pour la première fois que dans des cellules de mélanome agressifs, une fraction de la N-cadhérine est localisée dans les « lipid rafts » en association avec β-caténine, p 120ctn et α-caténine. Comme la présence de la N-cadhérine dans les « lipid rafts » ne contribue pas à la formation de jonction adhérentes, cette étude suggère une nouvelle fonction pour la N-cadhérine dans les « lipid rafts ». SUMMARY Human skin is a complex organ composed of two layers separated by a basement membrane: the epidermis and the dermis. In the basal layer of the epidermis, the melanin-producing cells of the skin, the melanocytes deliver melanin-containing melanosomes to keratinocytes, thereby protecting the epidermis and the dermis from the deleterious effects of ultraviolet light. Melanocytes physically interact with keratinocytes through E-cadherin-mediated adhesion. During malignant transformation into melanoma cells, melanocytes lose E-cadherin expression and concomitantly gain expression of N-cadherin, a phenomenon referred to as "cadherin switch". Loss of E-cadherin allows melanocytes to escape the regulatory effects of neighbouring keratinocytes, while gain of N-cadherin expression promotes migration, invasion and metastatic abilities of melanoma cells. In preliminary experiments, we found that a fraction of N-cadherin localized to specialized membrane microdomains enriched in cholesterol- and glycosphingolipid, called lipid rafts. One particular feature of lipid rafts is that they are rich in signalling molecules and they possibly modulate transmembrane signalling events. Moreover, recent reports suggested that a specialized type of rafts called caveolae might contribute to tumor progression. Based on the documented role of N-cadherin in melanoma progression and its presence in lipid rafts of melanoma cells, we raised the hypothesis that the association of N-cadherin with lipid rafts might be relevant to melanoma progression. The aim of this project was to characterize N-cadherin associated to lipid rafts during melanoma progression. Using human melanoma cell lines derived from melanoma at different stages of progression, we found that (1) N-cadherin is partly associated to lipid rafts in six cell lines derived from melanomas at late stages of progression and in experimental tumors, but not in two melanoma cell lines derived from early stages; (2) N-cadherin targeting to lipid rafts does not depend on its expression level; (3) E-cadherin is not localized in lipid rafts of a melanoma cell line that retained E-cadherin expression; (4) N-cadherin localization to lipid rafts is not modulated by the growth factors bFGF, IGF-I, and HRG1-β1, nor by MEK-, PKA-, Src family kinases-, and PI3K-mediated signalling events; (5) the association of N-cadherin with lipid rafts is not required for adherens junctions stability nor it is perturbed by adherens junctions disruption; (6) N-cadherin in lipid rafts is in complex with β-catenin, p 120ctm and α-catenin. In conclusion, this study provides original evidence that in aggressive melanoma cells a pool of N-cadherin is localized in lipid rafts in association with β-catenin, p 120 and α-catenin. The presence of N-cadherin in lipid rafts independently of its involvement in adherens junctions formation, suggests a possible new role for N-cadherin recruited to lipid rafts. Further studies investigating the biological meaning of this localization promise to uncover new properties of this molecule.

Selective changes in the neurofilament and microtubule cytoskeleton of NF-H/LacZ mice.

This study focused mainly on changes in the microtubule cytoskeleton in a transgenic mouse where beta-galactosidase fused to a truncated neurofilament subunit led to a decrease in neurofilament triplet protein expression and a loss in neurofilament assembly and abolished transport into neuronal processes in spinal cord and brain. Although all neurofilament subunits accumulated in neuronal cell bodies, our data suggest an increased solubility of all three subunits, rather than increased precipitation, and point to a perturbed filament assembly. In addition, reduced neurofilament phosphorylation may favor an increased filament degradation. The function of microtubules seemed largely unaffected, in that tubulin and microtubule-associated proteins (MAP) expression and their distribution were largely unchanged in transgenic animals. MAP1A was the only MAP with a reduced signal in spinal cord tissue, and differences in immunostaining in various brain regions corroborate a relationship between MAP1A and neurofilaments.

Binding of Lassa virus perturbs extracellular matrix-induced signal transduction via dystroglycan.

The arenavirus Lassa virus (LASV) causes a severe haemorrhagic fever with high mortality in man. The cellular receptor for LASV is dystroglycan (DG). DG is a ubiquitous receptor for extracellular matrix (ECM) proteins, which cooperates with β1 integrins to control cell-matrix interactions. Here, we investigated whether LASV binding to DG triggers signal transduction, mimicking the natural ligands. Engagement of DG by LASV resulted in the recruitment of the adaptor protein Grb2 and the protein kinase MEK1 by the cytoplasmic domain of DG without activating the MEK/ERK pathway, indicating assembly of an inactive signalling complex. LASV binding to cells however affected the activation of the MEK/ERK pathway via α6β1 integrins. The virus-induced perturbation of α6β1 integrin signalling critically depended on high-affinity LASV binding to DG and DG's cytoplasmic domain, indicating that LASV-receptor binding perturbed signalling cross-talk between DG and β1 integrins.