Batch effect confounding leads to strong bias in performance estimates obtained by cross-validation.

BACKGROUND: With the large amount of biological data that is currently publicly available, many investigators combine multiple data sets to increase the sample size and potentially also the power of their analyses. However, technical differences ("batch effects") as well as differences in sample composition between the data sets may significantly affect the ability to draw generalizable conclusions from such studies. FOCUS: The current study focuses on the construction of classifiers, and the use of cross-validation to estimate their performance. In particular, we investigate the impact of batch effects and differences in sample composition between batches on the accuracy of the classification performance estimate obtained via cross-validation. The focus on estimation bias is a main difference compared to previous studies, which have mostly focused on the predictive performance and how it relates to the presence of batch effects. DATA: We work on simulated data sets. To have realistic intensity distributions, we use real gene expression data as the basis for our simulation. Random samples from this expression matrix are selected and assigned to group 1 (e.g., 'control') or group 2 (e.g., 'treated'). We introduce batch effects and select some features to be differentially expressed between the two groups. We consider several scenarios for our study, most importantly different levels of confounding between groups and batch effects. METHODS: We focus on well-known classifiers: logistic regression, Support Vector Machines (SVM), k-nearest neighbors (kNN) and Random Forests (RF). Feature selection is performed with the Wilcoxon test or the lasso. Parameter tuning and feature selection, as well as the estimation of the prediction performance of each classifier, is performed within a nested cross-validation scheme. The estimated classification performance is then compared to what is obtained when applying the classifier to independent data.

Physicians' estimates of the 10-year cardiovascular risk in hypertensive patients

RESUME L'objectif de cette étude est d'évaluer comment de jeunes médecins en formation perçoivent le risque cardiovasculaire de leurs patients hypertendus en se basant sur les recommandations médicales (« guidelines ») et sur leur jugement clinique. Il s'agit d'une étude transversale observationnelle effectuée à la Policlinique Médicale Universitaire de Lausanne (PMU). 200 patients hypertendus ont été inclus dans l'étude ainsi qu'un groupe contrôle de 50 patients non hypertendus présentant au moins un facteur de risque cardiovasculaire. Nous avons comparé le risque cardiovasculaire à 10 ans calculé par un programme informatique basé sur l'équation de Framingham. L'équation a été adaptée pour les médecins par l'OMS-ISH au risque perçu, estimé cliniquement par les médecins. Les résultats de notre étude ont montrés que les médecins sous-estiment le risque cardiovasculaire à 10 ans de leurs patients, comparé au risque calculé selon l'équation de Framingham. La concordance entre les deux méthodes était de 39% pour les patients hypertendus et de 30% pour le groupe contrôle de patients non hypertendus. La sous-estimation du risque. cardiovasculaire pour les patients hypertendus était corrélée au fait qu'ils avaient une tension artérielle systolique stabilisée inférieure a 140 mmHg (OR=2.1 [1.1 ;4.1]). En conclusion, les résultats de cette étude montrent que les jeunes médecins en formation ont souvent une perception incorrecte du risque cardiovasculaire de leurs patients, avec une tendance à sous-estimer ce risque. Toutefois le risque calculé pourrait aussi être légèrement surestimé lorsqu'on applique l'équation de Framingham à la population suisse. Pour mettre en pratique une évaluation systématique des facteurs de risque en médecine de premier recours, un accent plus grand devrait être mis sur l'enseignement de l'évaluation du risque cardiovasculaire ainsi que sur la mise en oeuvre de programme pour l'amélioration de la qualité.

The Impact of Digitization on the Management of Administrative Procedures : The Case of Building Permits

Information and Communication Technologies provide public administrations new ways to meet their users' needs. At the same time, e-Government practices support the public sector in improving the quality of service provision and of its internal operations. In this paper we discuss the impacts of digitization on the management of administrative procedures. The theoretical framework and the research model that we will use in this study help us tackle the question of how digitization transforms administrative procedures as, for example, in terms of time and roles. The multiplicity of institutions involved in issuing building permits led us to consider this administrative procedure as a very interesting case study. An online survey was first addressed to Swiss civil servants to explore the field, and here we present some of its results. We are currently undertaking an in-depth case study of the building permit procedures in three Swiss Cantons, which we will also present in this paper. We will conclude with a discussion and the future steps of this project.

Sex-specific estimates of dispersal show female philopatry and male dispersal in a promiscuous amphibian, the alpine salamander (Salamandra atra).

Amphibians display wide variations in life-history traits and life cycles that should prove useful to explore the evolution of sex-biased dispersal, but quantitative data on sex-specific dispersal patterns are scarce. Here, we focused on Salamandra atra, an endemic alpine species showing peculiar life-history traits. Strictly terrestrial and viviparous, the species has a promiscuous mating system, and females reproduce only every 3 to 4 years. In the present study, we provide quantitative estimates of asymmetries in male vs. female dispersal using both field-based (mark-recapture) and genetic approaches (detection of sex-biased dispersal and estimates of migration rates based on the contrast in genetic structure across sexes and age classes). Our results revealed a high level of gene flow among populations, which stems exclusively from male dispersal. We hypothesize that philopatric females benefit from being familiar with their natal area for the acquisition and defence of an appropriate shelter, while male dispersal has been secondarily favoured by inbreeding avoidance. Together with other studies on amphibians, our results indicate that a species' mating system alone is a poor predictor of sex-linked differences in dispersal, in particular for promiscuous species. Further studies should focus more directly on the proximate forces that favour or limit dispersal to refine our understanding of the evolution of sex-biased dispersal in animals.

Closing the gap between T-cell life span estimates from stable isotope-labeling studies in mice and humans.

Quantitative knowledge of the turnover of different leukocyte populations is a key to our understanding of immune function in health and disease. Much progress has been made thanks to the introduction of stable isotope labeling, the state-of-the-art technique for in vivo quantification of cellular life spans. Yet, even leukocyte life span estimates on the basis of stable isotope labeling can vary up to 10-fold among laboratories. We investigated whether these differences could be the result of variances in the length of the labeling period among studies. To this end, we performed deuterated water-labeling experiments in mice, in which only the length of label administration was varied. The resulting life span estimates were indeed dependent on the length of the labeling period when the data were analyzed using a commonly used single-exponential model. We show that multiexponential models provide the necessary tool to obtain life span estimates that are independent of the length of the labeling period. Use of a multiexponential model enabled us to reduce the gap between human T-cell life span estimates from 2 previously published labeling studies. This provides an important step toward unambiguous understanding of leukocyte turnover in health and disease.

Stereologic estimates of total spinophilin-immunoreactive spine number in area 9 and the CA1 field: Relationship with the progression of Alzheimer's disease

The loss of presynaptic markers is thought to represent a strong pathologic correlate of cognitive decline in Alzheimer's disease (AD). Spinophilin is a postsynaptic marker mainly located to the heads of dendritic spines. We assessed total numbers of spinophilin-immunoreactive puncta. in the CA I and CA3 fields of hippocampus and area 9 in 18 elderly individuals with various degrees of cognitive decline. The decrease in spinophilin-immunoreactivity was significantly related to both Braak neurofibrillary tangle (NFT) staging and clinical severity but not A beta deposition staging. The total number of spinophilin-immunoreactive puncta in CA I field and area 9 were significantly related to MMSE scores and predicted 23.5 and 61.9% of its variability. The relationship between total number of spinophilin-immunoreactive puncta in CA I field and MMSE scores did not persist when adjusting for Braak NFT staging. In contrast, the total number of spinophilin-immunoreactive puncta in area 9 was still significantly related to the cognitive outcome explaining an extra 9.6% of MMSE and 25.6% of the Clinical Dementia Rating scores variability. Our data suggest that neocortical dendritic spine loss is an independent parameter to consider in AD clinicopathologic correlations.

Representational pseudoneglect and reference points both influence geographic location estimates

Our mental representation of the world is far from objective. For example, western Canadians estimate the locations of North American cities to be too far to the west. This bias could be due to a reference point effect, in which people estimate more space between places close to them than far from them, or to representational pseudoneglect, in which neurologically intact individuals favor the left side of space when asked to image a scene.We tested whether either or both of these biases influence the geographic world representations of neurologically intact young adults from Edmonton and Ottawa, which are in western and eastern Canada, respectively. Individuals were asked to locate NorthAmerican cities on a two-dimensional grid. Both groups revealed effects of representational pseudoneglect in this novel paradigm, but they also each exhibited reference point effects. These results inform theories in both cognitive psychology and neuroscience.

Reproducibility of species lists, visual cover estimates and frequency methods for recording high mountain vegetation

Question: When multiple observers record the same spatial units of alpine vegetation, how much variation is there in the records and what are the consequences of this variation for monitoring schemes to detect change? Location: One test summit in Switzerland (Alps) and one test summit in Scotland (Cairngorm Mountains). Method: Eight observers used the GLORIA protocols for species composition and visual cover estimates in percent on large summit sections (>100 m2) and species composition and frequency in nested quadrats (1 m2). Results: The multiple records from the same spatial unit for species composition and species cover showed considerable variation in the two countries. Estimates of pseudoturnover of composition and coefficients of variation of cover estimates for vascular plant species in 1m x 1m quadrats showed less variation than in previously published reports whereas our results in larger sections were broadly in line with previous reports. In Scotland, estimates for bryophytes and lichens were more variable than for vascular plants. Conclusions: Statistical power calculations indicated that, unless large numbers of plots were used, changes in cover or frequency were only likely to be detected for abundant species (exceeding 10% cover) or if relative changes were large (50% or more). Lower variation could be reached with the point methods and with larger numbers of small plots. However, as summits often strongly differ from each other, supplementary summits cannot be considered as a way of increasing statistical power without introducing a supplementary component of variance into the analysis and hence the power calculations.

One-year prevalence of low back pain in two Swiss regions: estimates from the population participating in the 1992-1993 MONICA project.

STUDY DESIGN: A cross-sectional survey was performed. OBJECTIVE: To estimate the extent of low back pain as a public health problem. SUMMARY OF BACKGROUND DATA: Health surveys converge on very high estimates of low back pain in general populations, but few studies have included severity criteria in their definition and conclusions. Because it is unlikely that interventions will influence the prevalence of minimal and infrequent symptoms, greater attention should be paid to characteristics of low back pain that indicate some impact on the life of survey respondents. METHODS: Two regions participated in the MONICA (MONitoring of trends and determinants in CArdiovascular disease) project in Switzerland. Participants randomly selected from the general population completed a standard self-administered questionnaire on cardiovascular risk factors. A special section on low back pain was added in the third (1992-1993) MONICA survey and completed by 3227 participants. RESULTS: A regional difference found in the 12-month prevalence rate disappeared with the inclusion of severity criteria. Low back pain over more than seven cumulated days was reported among men by 20.2% (age range, 25-34 years) to 28.5% (age range, 65-74 years), respectively, among women by 31.1% to 38.5%. Similar rates of reduction in activity (professional, housekeeping, and leisure time) and medical consultation (conventional and nonconventional) motivated by low back pain characterized the two participating regions. The cumulative duration of pain was related to all the indicators showing the impact of low back pain on everyday life. CONCLUSIONS: Determining the cumulative duration of low back pain over the preceding year is a straightforward task, and a cutoff at 1 week seems appropriate for distinguishing between low- and high-impact low back pain.

Building hippocampal circuits to learn and remember: insights into the development of human memory

The hippocampal formation is essential for the processing of episodic memories for autobiographical events that happen in unique spatiotemporal contexts. Interestingly, before 2 years of age, children are unable to form or store episodic memories for recall later in life, a phenomenon known as infantile amnesia. From 2 to 7 years of age, there are fewer memories than predicted based on a forgetting function alone, a phenomenon known as childhood amnesia. Here, we discuss the postnatal maturation of the primate hippocampal formation with the goal of characterizing the development of the neurobiological substrates thought to subserve the emergence of episodic memory. Distinct regions, layers and cells of the hippocampal formation exhibit different profiles of structural and molecular development during early postnatal life. The protracted period of neuronal addition and maturation in the dentate gyrus is accompanied by the late maturation of specific layers in different hippocampal regions that are located downstream from the dentate gyrus, particularly CA3. In contrast, distinct layers in several hippocampal regions, particularly CA1, which receive direct projections from the entorhinal cortex, exhibit an early maturation. In addition, hippocampal regions that are more highly interconnected with subcortical structures, including the subiculum, presubiculum, parasubiculum and CA2, mature even earlier. These findings, together with our studies of the development of human spatial memory, support the hypothesis that the differential maturation of distinct hippocampal circuits might underlie the differential emergence of specific "hippocampus-dependent" memory processes, culminating in the emergence of episodic memory concomitant with the maturation of all hippocampal circuits.

Assessing inter-rater reliability when the raters are fixed: two concepts and two estimates.

Intraclass correlation (ICC) is an established tool to assess inter-rater reliability. In a seminal paper published in 1979, Shrout and Fleiss considered three statistical models for inter-rater reliability data with a balanced design. In their first two models, an infinite population of raters was considered, whereas in their third model, the raters in the sample were considered to be the whole population of raters. In the present paper, we show that the two distinct estimates of ICC developed for the first two models can both be applied to the third model and we discuss their different interpretations in this context.