Jean-Yves Dauxois

Ultrastructural localization of rRNA shows defective nuclear export of preribosomes in mutants of the Nup82p complex.

To study the nuclear export of preribosomes, ribosomal RNAs were detected by in situ hybridization using fluorescence and EM, in the yeast Saccharomyces cerevisiae. In wild-type cells, semiquantitative analysis shows that the distributions of pre-40S and pre-60S particles in the nucleolus and the nucleoplasm are distinct, indicating uncoordinated transport of the two subunits within the nucleus. In cells defective for the activity of the GTPase Gsp1p/Ran, ribosomal precursors accumulate in the whole nucleus. This phenotype is reproduced with pre-60S particles in cells defective in pre-rRNA processing, whereas pre-40S particles only accumulate in the nucleolus, suggesting a tight control of the exit of the small subunit from the nucleolus. Examination of nucleoporin mutants reveals that preribosome nuclear export requires the Nup82p–Nup159p–Nsp1p complex. In contrast, mutations in the nucleoporins forming the Nup84p complex yield very mild or no nuclear accumulation of preribosome. Interestingly, domains of Nup159p required for mRNP trafficking are not necessary for preribosome export. Furthermore, the RNA helicase Dbp5p and the protein Gle1p, which interact with Nup159p and are involved in mRNP trafficking, are dispensable for ribosomal transport. Thus, the Nup82p–Nup159p–Nsp1p nucleoporin complex is part of the nuclear export pathways of preribosomes and mRNPs, but with distinct functions in these two processes.

On testing the proportionality of two cumulative incidence functions in a competing risks setup

A large sample test of proportionality of two cumulative incidence functions is developed for randomly right censored competing risks data without assuming that the K ≥ 2 risks are independent. The test is tailored to detecting monotonicity of the ratio of the two cumulative incidence functions and the test statistic is proved to be asymptotically normally distributed. In addition, it is shown that the proposed test can be readily adapted to test whether the censoring random variable and observable survival time have proportional hazards. The procedures are illustrated through application to two data sets well known in the survival analysis literature.

Bayesian inference for linear growth birth and death processes

In this paper, we are interested in Bayesian inference for two kinds of birth and death Markov processes those of linear growth with immigration (see Karlin and Taylor (A First Course in Stochastic Processes, 2nd Edition, Academic Press, New York)) and linear growth with limited population (see Kleinrock (Queueing Systems, Volume 1: Theory, Wiley, New York)). The Gauss-hypergeometric distribution (Armero and Bayarri, The Statistician 43 (1994b) 139) is used to introduce conjugate priors. Bayesian estimates of some measures of performance are obtained.

A bayesian choice between poisson, binomial and negative binomial models

In this paper, we propose a Bayesian method for modelling count data by Poisson, binomial or negative binomial distributions. These three distributions have in common that the variance is, at most, a quadratic function of the mean. We use prior distributions on the variance function coefficients to consider simultaneously the three possible models and decide which one fits the data better. This approach sheds new light on the analysis of the Sibship data (Sokal and Rohlf, 1987). The Jeffreys-Lindley paradox is discussed through some illustrations.

A new method for proving weak convergence results applied to nonparametric estimators in survival analysis

Using the limit theorem for stochastic integral obtained by Jakubowski et al. (Probab. Theory Related Fields 81 (1989) 111-137), we introduce in this paper a new method for proving weak convergence results of empirical processes by a martingale method which allows discontinuities for the underlying distribution. This is applied to Nelson-Aalen and Kaplan-Meier processes. We also prove that the same conclusion can be drawn for Hjorts nonparametric Bayes estimators of the cumulative distribution function and cumulative hazard rate.

Estimating time-to-onset of adverse drug reactions from spontaneous reporting databases

BackgroundAnalyzing time-to-onset of adverse drug reactions from treatment exposure contributes to meeting pharmacovigilance objectives, i.e. identification and prevention. Post-marketing data are available from reporting systems. Times-to-onset from such databases are right-truncated because some patients who were exposed to the drug and who will eventually develop the adverse drug reaction may do it after the time of analysis and thus are not included in the data. Acknowledgment of the developments adapted to right-truncated data is not widespread and these methods have never been used in pharmacovigilance. We assess the use of appropriate methods as well as the consequences of not taking right truncation into account (naive approach) on parametric maximum likelihood estimation of time-to-onset distribution.MethodsBoth approaches, naive or taking right truncation into account, were compared with a simulation study. We used twelve scenarios for the exponential distribution and twenty-four for the Weibull and log-logistic distributions. These scenarios are defined by a set of parameters: the parameters of the time-to-onset distribution, the probability of this distribution falling within an observable values interval and the sample size. An application to reported lymphoma after anti TNF- α treatment from the French pharmacovigilance is presented.ResultsThe simulation study shows that the bias and the mean squared error might in some instances be unacceptably large when right truncation is not considered while the truncation-based estimator shows always better and often satisfactory performances and the gap may be large. For the real dataset, the estimated expected time-to-onset leads to a minimum difference of 58 weeks between both approaches, which is not negligible. This difference is obtained for the Weibull model, under which the estimated probability of this distribution falling within an observable values interval is not far from 1.ConclusionsIt is necessary to take right truncation into account for estimating time-to-onset of adverse drug reactions from spontaneous reporting databases.

Parametric Bootstrap Goodness-of-Fit Tests for Imperfect Maintenance Models

The simultaneous modeling of ageing and maintenance efficiency of repairable systems is a major issue in reliability. Many imperfect maintenance models have been proposed. To analyze a dataset, it is necessary to check whether these models are adapted or not. In this paper, we propose a general methodology for testing the goodness of fit of any kind of imperfect maintenance model. Two families of tests are presented, based respectively on martingale residuals and probability integral transforms. The quantiles of the test statistics distributions under the null hypothesis are computed with parametric bootstrap methods. An extensive simulation study is provided, from which we recommend the use of two tests in practice, one from each family. Finally, the tests are applied to several real datasets.

Testing relative risk under random censoring

We consider the problem of testing H0 : F1=F2 on the basis of two independent random samples of randomly right censored survival times. Tests for H0 are developed when the alternatives of primary interest are of the type is increasing (decreasing) or H* : F1/F2 is increasing (decreasing). Asymptotic normality of the test statistic is proved by counting process techniques and the testing procedures are illustrated through application to a well-known data set in the survival analysis literature.

On the Parametric Maximum Likelihood Estimator for Independent but Non-identically Distributed Observations with Application to Truncated Data

We investigate the parametric maximum likelihood estimator for truncated data when the truncation value is different according to the observed individual or item. We extend Lehmanns proof (1983) of the asymptotic properties of the parametric maximum likelihood estimator in the case of independent non-identically distributed observations. Two cases are considered: either the number of distinct probability distribution functions that can be observed in the population from which the sample comes from is finite or this number is infinite. Sufficient conditions for consistency and asymptotic normality are provided for both cases.

Semiparametric inference of competing risks data with additive hazards and missing cause of failure under MCAR or MAR assumptions

In this paper, we consider a semiparametric model for lifetime data with competing risks and missing causes of death. We assume that an additive hazards model holds for each cause-specific hazard rate function and that a random right censoring occurs. Our goal is to estimate the regression parameters as well as the functional parameters such as the baseline and cause-specific cumulative hazard rate functions / cumulative incidence functions. We first introduce preliminary estimators of the unknown (Euclidean and functional) parameters when cause of death indicators are missing completely at random (MCAR). These estimators are obtained using the observations with known cause of failure. The advantage of considering the MCAR model is that the information given by the observed lifetimes with unknown failure cause can be used to improve the preliminary estimates in order to attain an asymptotic optimality criterion. This is the main purpose of our work. However, since it is often more realistic to consider a missing at random (MAR) mechanism, we also derive estimators of the regression and functional parameters under the MAR model. We study the large sample properties of our estimators through martingales and empirical process techniques. We also provide a simulation study to compare the behavior of our three types of estimators under the different mechanisms of missingness. It is shown that our improved estimators under MCAR assumption are quite robust if only the MAR assumption holds. Finally, three illustrations on real datasets are also given.