Robust Estimates of the Negative Binomial Model

We consider robust parametric procedures for univariate discrete distributions, focusing on the negative binomial model. The procedures are based on three steps: ?First, a very robust, but possibly inefficient, estimate of the model parameters is computed. ?Second, this initial model is used to identify outliers, which are then removed from the sample. ?Third, a corrected maximum likelihood estimator is computed with the remaining observations. The final estimate inherits the breakdown point (bdp) of the initial one and its efficiency can be significantly higher. Analogous procedures were proposed in [1], [2], [5] for the continuous case. A comparison of the asymptotic bias of various estimates under point contamination points out the minimum Neyman's chi-squared disparity estimate as a good choice for the initial step. Various minimum disparity estimators were explored by Lindsay [4], who showed that the minimum Neyman's chi-squared estimate has a 50% bdp under point contamination; in addition, it is asymptotically fully efficient at the model. However, the finite sample efficiency of this estimate under the uncontaminated negative binomial model is usually much lower than 100% and the bias can be strong. We show that its performance can then be greatly improved using the three step procedure outlined above. In addition, we compare the final estimate with the procedure described in

Clinical and course characteristics of depression and all-cause mortality: A prospective population-based study.

BACKGROUND: Given the large heterogeneity of depressive disorders (DD), studying depression characteristics according to clinical manifestations and course is a more promising approach than studying depression as a whole. The purpose of this study was to determine the association between clinical and course characteristics of DD and incident all-cause mortality. METHODS: CoLaus|PsyCoLaus is a prospective cohort study (mean follow-up duration=5.2 years) including 35-66 year-old randomly selected residents of an urban area in Switzerland. A total of 3668 subjects (mean age 50.9 years, 53.0% women) underwent physical and psychiatric baseline evaluations and had a known vital status at follow-up (98.8% of the baseline sample). Clinical (diagnostic severity, atypical features) and course characteristics (recency, recurrence, duration, onset) of DD according to the DSM-5 were elicited using a semi-structured interview. RESULTS: Compared to participants who had never experienced DD, participants with current but not remitted DD were more than three times as likely to die (Hazard Ratio: 3.2, 95% CI: 1.1-10.0) after adjustment for socio-demographic and lifestyle characteristics, comorbid anxiety disorders, antidepressant use, and cardiovascular risk factors and diseases. There was no evidence for associations between other depression characteristics and all-cause mortality. LIMITATIONS: The small proportion of deceased subjects impeded statistical analyses of cause-specific mortality. CONCLUSIONS: A current but not remitted DD is a strong predictor of all-cause mortality, independently of cardiovascular or lifestyle factors, which suggests that the effect of depression on mortality diminishes after remission and further emphasizes the need to adequately treat current depressive episodes.

An automated microreactor for semi-continuous biosensor measurements.

Living bacteria or yeast cells are frequently used as bioreporters for the detection of specific chemical analytes or conditions of sample toxicity. In particular, bacteria or yeast equipped with synthetic gene circuitry that allows the production of a reliable non-cognate signal (e.g., fluorescent protein or bioluminescence) in response to a defined target make robust and flexible analytical platforms. We report here how bacterial cells expressing a fluorescence reporter ("bactosensors"), which are mostly used for batch sample analysis, can be deployed for automated semi-continuous target analysis in a single concise biochip. Escherichia coli-based bactosensor cells were continuously grown in a 13 or 50 nanoliter-volume reactor on a two-layered polydimethylsiloxane-on-glass microfluidic chip. Physiologically active cells were directed from the nl-reactor to a dedicated sample exposure area, where they were concentrated and reacted in 40 minutes with the target chemical by localized emission of the fluorescent reporter signal. We demonstrate the functioning of the bactosensor-chip by the automated detection of 50 μgarsenite-As l(-1) in water on consecutive days and after a one-week constant operation. Best induction of the bactosensors of 6-9-fold to 50 μg l(-1) was found at an apparent dilution rate of 0.12 h(-1) in the 50 nl microreactor. The bactosensor chip principle could be widely applicable to construct automated monitoring devices for a variety of targets in different environments.