Transfer Learning in Multiple Hypothesis Testing.

In this investigation, a synthesis of Convolutional Neural Networks (CNNs) and Bayesian inference is presented, leading to a novel approach to the problem of Multiple Hypothesis Testing (MHT). Diverging from traditional paradigms, this study introduces a sequence-based uncalibrated Bayes factor approach to test many hypotheses using the same family of sampling parametric models. A two-step methodology is employed: initially, a learning phase is conducted utilizing simulated datasets encompassing a wide spectrum of null and alternative hypotheses, followed by a transfer phase applying this fitted model to real-world experimental sequences. The outcome is a CNN model capable of navigating the complex domain of MHT with improved precision over traditional methods, also demonstrating robustness under varying conditions, including the number of true nulls and dependencies between tests. Although indications of empirical evaluations are presented and show that the methodology will prove useful, more work is required to provide a full evaluation from a theoretical perspective. The potential of this innovative approach is further illustrated within the critical domain of genomics. Although formal proof of the consistency of the model remains elusive due to the inherent complexity of the algorithms, this paper also provides some theoretical insights and advocates for continued exploration of this methodology.

Model uncertainty quantification in Cox regression.

We consider covariate selection and the ensuing model uncertainty aspects in the context of Cox regression. The perspective we take is probabilistic, and we handle it within a Bayesian framework. One of the critical elements in variable/model selection is choosing a suitable prior for model parameters. Here, we derive the so-called conventional prior approach and propose a comprehensive implementation that results in an automatic procedure. Our simulation studies and real applications show improvements over existing literature. For the sake of reproducibility but also for its intrinsic interest for practitioners, a web application requiring minimum statistical knowledge implements the proposed approach.

The relationship among the age at menarche, anthropometric characteristics, and socio-economic factors in Bengali girls from Kolkata, India.

OBJECTIVES: The aim of the study was to measure the causal effect of selected socio-economic factors and anthropometrical characteristics on the menarche occurrence. METHODS: The sample consisted of 2195 Bengali girls (aged 7-21) from middle-class families, from Kolkata city, India. The age at menarche was recorded from the retrospective data and status quo method. The causal effect of anthropometric and socio-economic variables on menarche occurrence was estimated by the nonparametrical analysis of survival probability (survival random forest). RESULTS: In the examined cohort menarche occurred, on average, at 11.8 years of age. The probability of menarche occurrence amplified with the increasing values of factors such as body mass index, height-for-age z-scores, number of family members, household rooms, and toilets, but decreased when expenditures increased. The relation maintained a similar pattern of causal effect with girls' age. CONCLUSIONS: A complex pattern of relationship among sexual development, physique, and socio-economic characteristics was defined. The tendency toward early menarche, along with the observed causal relationships indicate that the analyzed sample is nearing the characteristics and standards of living noted in other middle and even high-income countries in the world.

A Dirichlet Process Prior Approach for Covariate Selection.

The variable selection problem in general, and specifically for the ordinary linear regression model, is considered in the setup in which the number of covariates is large enough to prevent the exploration of all possible models. In this context, Gibbs-sampling is needed to perform stochastic model exploration to estimate, for instance, the model inclusion probability. We show that under a Bayesian non-parametric prior model for analyzing Gibbs-sampling output, the usual empirical estimator is just the asymptotic version of the expected posterior inclusion probability given the simulation output from Gibbs-sampling. Other posterior conditional estimators of inclusion probabilities can also be considered as related to the latent probabilities distributions on the model space which can be sampled given the observed Gibbs-sampling output. This paper will also compare, in this large model space setup the conventional prior approach against the non-local prior approach used to define the Bayes Factors for model selection. The approach is exposed along with simulation samples and also an application of modeling the Travel and Tourism factors all over the world.

Nutritional status of Ugandan school-children: The effect of age imprecision.

OBJECTIVES: To analyze the nutritional status of Ugandan school-children in a cross-sectional and longitudinal perspective, considering the effect of age imprecision. MATERIALS AND METHODS: Anthropometric measurements of 831 school-children (381 males and 450 females) were analyzed. A subsample of 246 children was measured in July 2014 and 2015. Stunting (based on height-for-age Z-scores), underweight (weight-for-age), and thinness (body mass index-for-age) prevalence were calculated. Three different ages were used: declared (from schools registers), attributed (based on multiple information sources), and bootstrap (from 10,000 replicates). Significant differences among malnutrition prevalence calculated with different ages and in different groups were assessed by means of bootstrap analysis. Longitudinal analysis was conducted using a paired t test. RESULTS: The mean prevalence of malnutrition calculated with declared, attributed, or bootstrap ages were very similar: stunting (11.9-12.7); underweight (5.4-5.9); thinness (3.3-3.7); and obesity (0.7). Undernutrition was more prevalent among older children, while obesity was mostly associated with young age. Obesity was equally distributed among sexes, while undernutrition was more prevalent among females of up to 10 years of age and males above 10 years. The longitudinal analysis indicated a reduction in underweight and thinness, and an increase in stunting, especially among older children. DISCUSSION: Age imprecision did not significantly affect malnutrition estimates. Despite the decline in the prevalence of thinness and underweight observed over a 1-year period, undernutrition persists, with an observed rise in stunting. On the other hand, obesity is starting to appear. Public health efforts are required to eliminate stunting and address the emerging burden of obesity.

P-value calibration in multiple hypotheses testing.

As p-values are the most common measures of evidence against a hypothesis, their calibration with respect to null hypothesis conditional probability is important in order to match frequentist unconditional inference with the Bayesian ones. The Selke, Bayarri and Berger calibration is one of the most popular attempts to obtain such a calibration. This relies on the theoretical sampling null distribution of p-values, which is the well-known Uniform(0,1), but arising only for specific sampling models. We generalize this calibration by considering a sampling null distribution estimated from the data. It is possible to obtain such an empirical null distribution, for instance, in the context of multiple testing in which many p-values come from the null model. Such a context is purely instrumental for the purposes of p-value calibration, and multiple testing still needs to be considered with appropriate techniques. The new calibration proposed here still remains a simple analytic formula like the original one under the Uniform(0,1) and basically provides a stronger interpretation framework for the widely used p-value. Copyright © 2017 John Wiley & Sons, Ltd.

Birth registration and child undernutrition in sub-Saharan Africa.

OBJECTIVE: In many countries of the world millions of people are not registered at birth. However, in order to assess children's nutritional status it is necessary to have an exact knowledge of their age. In the present paper we discuss the effects of insufficient or imprecise age data on estimates of undernutrition prevalence. DESIGN: Birth registration rates and levels of stunting, underweight and wasting were retrieved from Multiple Indicator Cluster Surveys and Demographic and Health Surveys of thirty-seven sub-Saharan African countries, considering the subdivision in wealth quintiles. The composition of the cross-sectional sample used for nutritional evaluation was analysed using a permutation test. Logistic regression was applied to analyse the relationship between birth registration and undernutrition. The 95 % probability intervals and Student's t test were used to evaluate the effect of age bias and error. RESULTS: Heterogeneous sampling designs were detected among countries, with different percentages of children selected for anthropometry. Further, registered children were slightly more represented within samples used for nutritional analysis than in the total sample. A negative relationship between birth registration and undernutrition was recognized, with registered children showing a better nutritional status than unregistered ones, even within each wealth quintile. The over- or underestimation of undernutrition in the case of systematic over- or underestimation of age, respectively, the latter being more probable, was quantified up to 28 %. Age imprecision was shown to slightly overestimate undernutrition. CONCLUSIONS: Selection bias towards registered children and underestimation of children's age can lead to an underestimation of the prevalence of undernutrition.

Comparison of objective Bayes factors for variable selection in parametric regression models for survival analysis.

This paper considers the problem of selecting a set of regressors when the response variable is distributed according to a specified parametric model and observations are censored. Under a Bayesian perspective, the most widely used tools are Bayes factors (BFs), which are undefined when improper priors are used. In order to overcome this issue, fractional (FBF) and intrinsic (IBF) BFs have become common tools for model selection. Both depend on the size, Nt , of a minimal training sample (MTS), while the IBF also depends on the specific MTS used. In the case of regression with censored data, the definition of an MTS is problematic because only uncensored data allow to turn the improper prior into a proper posterior and also because full exploration of the space of the MTSs, which includes also censored observations, is needed to avoid bias in model selection. To address this concern, a sequential MTS was proposed, but it has the drawback of an increase of the number of possible MTSs as Nt becomes random. For this reason, we explore the behaviour of the FBF, contextualizing its definition to censored data. We show that these are consistent, providing also the corresponding fractional prior. Finally, a large simulation study and an application to real data are used to compare IBF, FBF and the well-known Bayesian information criterion.

Food drugs as drivers of therapeutic knowledge and the role of chemosensory qualities.

ETHNOPHARMACOLOGICAL RELEVANCE: Chemosensory qualities of botanical drugs are important cues for anticipating physiologic consequences. Whether a botanical drug is used for both, food and medicine, or only as medicine depends on taste preferences, nutritional content, cultural background, and the individual and overall epidemiological context. MATERIAL AND METHODS: We subjected 540 botanical drugs described in De Materia Medica having at least one oral medical application to a tasting panel. The 540 drugs were grouped into those only used for medicine (388) and those also used for food (152). The associations with chemosensory qualities and therapeutic indications were compared across the two groups. We considered 22 experimentally assessed chemosensory qualities and 39 categories of therapeutic use groups. We wanted to know, 1): which chemosensory qualities increase the probability of an orally applied botanical drug to be also used for food ? 2): which chemosensory qualities augment the probability of an orally applied botanical drug to be only used for medicine? and 3): whether there are differences in therapeutic indications between orally applied botanical drugs also used for food (food drugs) and botanical drugs applied exclusively for medicinal purposes (non-food drugs) and, if yes, how the differences can be explained. RESULTS: Chemosensory qualities augmenting the probability of an orally applied botanical drug to be also used for food were sweet, starchy, salty, burning/hot, fruity, nutty, and cooling. Therapeutics used for diarrhoea, as libido modulators, purgatives, laxatives, for expelling parasites, breast and lactation and increasing diuresis, were preferentially sourced from food drugs while drugs used for liver and jaundice, vaginal discharge and humoral management showed significant negative associations with food dugs in ancient Greek-Roman materia medica. CONCLUSION: Therapeutics used for ailments of body organs involved in the digestion of food and the excretion of waste products showed a tendency to be sourced from food drugs. Arguably, the daily consumption of food offered the possibility for observing post-prandial physiologic and pharmacologic effects which led to a high therapeutic versatility of food drugs and the possibility to understand benefits of taste and flavour qualities. The difference in chemosensory qualities between food drugs and non-food drugs is demarcating the organoleptic requirements of food rather than that of medicine.

Usability of classic and specific bioelectrical impedance vector analysis in measuring body composition of children.

In this study, we aimed to analyse the relationship between body composition and bioelectrical variables in children and adolescents. The sample was composed of 6801 individuals (4035 males; 2766 females) aged 8-20 years included in the National Health and Nutrition Examination Survey (NHANES) years 1999-2004. Classic and specific bioelectrical impedance vector analysis (BIVA) were applied and compared with dual-energy X-ray absorptiometry (DXA) for the evaluation of fat mass (FM) and fat-free mass (FFM), and bioimpedance spectroscopy (BIS) for the evaluation of intra-cellular water (ICW), extra-cellular water (ECW), and total body water (TBW). Fat-free mass index (FFMI) was calculated. Spearman's correlation, regression, and depth-depth analyses were applied. The evaluation of body composition with BIVA agreed well with that of DXA or BIS, independently of sex, age, and ethnicity: classic BIVA was mostly sensitive to differences in TBW, ECW/ICW, whereas specific BIVA to differences in %FM, FFMI, and ECW/ICW. The depth-depth analysis confirmed the associations of classic BIVA (coeff. 0.500, p < 0.001), and specific BIVA (coeff. 0.512, p < 0.001), also considering the significant effect of age (p < 0.001). In classic BIVA the association was slightly stronger in females (by 0.03, p = 0.042) and among Blacks (0.06, p = 0.002), whereas in specific BIVA it was stronger by 0.06 (p < 0.001) in females and similar among ethnic groups. The combined use of the two BIVA approaches represents a valuable tool for complete evaluation of body composition in growth studies, for the prevention and monitoring of malnutrition, and the monitoring of the performance in young athletes.

A strategy analysis for genetic association studies with known inbreeding.

BACKGROUND: Association studies consist in identifying the genetic variants which are related to a specific disease through the use of statistical multiple hypothesis testing or segregation analysis in pedigrees. This type of studies has been very successful in the case of Mendelian monogenic disorders while it has been less successful in identifying genetic variants related to complex diseases where the insurgence depends on the interactions between different genes and the environment. The current technology allows to genotype more than a million of markers and this number has been rapidly increasing in the last years with the imputation based on templates sets and whole genome sequencing. This type of data introduces a great amount of noise in the statistical analysis and usually requires a great number of samples. Current methods seldom take into account gene-gene and gene-environment interactions which are fundamental especially in complex diseases. In this paper we propose to use a non-parametric additive model to detect the genetic variants related to diseases which accounts for interactions of unknown order. Although this is not new to the current literature, we show that in an isolated population, where the most related subjects share also most of their genetic code, the use of additive models may be improved if the available genealogical tree is taken into account. Specifically, we form a sample of cases and controls with the highest inbreeding by means of the Hungarian method, and estimate the set of genes/environmental variables, associated with the disease, by means of Random Forest. RESULTS: We have evidence, from statistical theory, simulations and two applications, that we build a suitable procedure to eliminate stratification between cases and controls and that it also has enough precision in identifying genetic variants responsible for a disease. This procedure has been successfully used for the beta-thalassemia, which is a well known Mendelian disease, and also to the common asthma where we have identified candidate genes that underlie to the susceptibility of the asthma. Some of such candidate genes have been also found related to common asthma in the current literature. CONCLUSIONS: The data analysis approach, based on selecting the most related cases and controls along with the Random Forest model, is a powerful tool for detecting genetic variants associated to a disease in isolated populations. Moreover, this method provides also a prediction model that has accuracy in estimating the unknown disease status and that can be generally used to build kit tests for a wide class of Mendelian diseases.

Segmental body composition estimated by specific BIVA and dual-energy X-ray absorptiometry.

AIMS: The aim of this study was to analyse the association between specific bioelectric impedance vector analysis (BIVA) and dual-energy X-ray absorptiometry (DXA) to assess segmental body composition using DXA as the reference technique. METHODS: The sample comprised 50 young active students who practised or played different sports (25 men, age: 24.37 ± 4.79 y; 25 women, age: 24.32 ± 4.43 y) from the National Institute of Physical Education of Catalonia (INEFC). Anthropometric data (height, weight, arm, waist, and calf circumferences) and bioelectrical measurements (R, ohm; Xc, ohm) were recorded. Body composition was analysed with specific BIVA. DXA was used as the reference method to assess body composition of the whole-body, the trunk, and the limbs. The percentage of fat mass (%FMDXA) and fat-free mass index (FFMIDXA = FFM/length2) were calculated. The agreement between specific BIVA and DXA was evaluated by a depth-depth analysis, two-way ANOVA, and Pearson's correlations. RESULTS: The depth-depth analysis showed a good agreement between DXA and BIVA (F = 14.89, p < 0.001) in both sexes and all body segments. Specific vector length (Zsp; i.e. indicative of %FM) was correlated with %FMDXA in the whole body and all body segments, and the phase angle was correlated with FFMIDXA, with he trunk in women as the only exception. Specific BIVA demonstrated to balance the effect of body size on bioelectrical measurements in both whole and segmental approaches. CONCLUSIONS: Segmental specific BIVA and DXA provided a consistent evaluation of body composition in both sexes, of the whole body and each body segment. The indices %FM and FFMI obtained with DXA were correlated to vector length and phase angle in each segment, respectively. Specific BIVA represents a promising technique for monitoring segmental body composition changes in sport science and clinical applications.

Antidepressant imipramine induces human astrocytes to differentiate into cells with neuronal phenotype.

Several recent studies have expanded our conception of the role of astrocytes in neurogenesis, proposing that these cells may contribute to this phenomenon not only as a source of trophic substances, but also as stem cells themselves. We recently observed in vitro that human mature astrocytes can be induced to differentiate into cells with a neuronal phenotype. Antidepressant drugs have been shown to increase neurogenesis in the adult rodent hippocampus. In order to better understand the role of astroglia in antidepressant-induced neurogenesis, primary astrocyte cultures were treated with the antidepressant imipramine. Cell morphology was rapidly modified by treatment. In fact, whereas untreated astrocytes showed large, flat morphology, after a few hours of treatment cells exhibited a round-shaped cell body with long, thin processes. The expression of neuronal markers was analysed by immunocytochemistry, Western Blot and RT-PCR at different treatment times. Results showed an increase in neuronal markers such as neurofilament and neuron-specific enolase (NSE), whereas glial fibrillary acidic protein (GFAP) and nestin expression were not significantly modified by treatment. Similar results were obtained with fluoxetine and venlafaxine. Hes1 mRNA significantly increased after 2 h of treatment, suggesting involvement of this transcription factor in this process. These results confirm the role of astrocytes in neurogenesis and suggest that these cells may represent one of the targets of antidepressants.

A Markov chain representation of the multiple testing problem.

The problem of multiple hypothesis testing can be represented as a Markov process where a new alternative hypothesis is accepted in accordance with its relative evidence to the currently accepted one. This virtual and not formally observed process provides the most probable set of non null hypotheses given the data; it plays the same role as Markov Chain Monte Carlo in approximating a posterior distribution. To apply this representation and obtain the posterior probabilities over all alternative hypotheses, it is enough to have, for each test, barely defined Bayes Factors, e.g. Bayes Factors obtained up to an unknown constant. Such Bayes Factors may either arise from using default and improper priors or from calibrating p-values with respect to their corresponding Bayes Factor lower bound. Both sources of evidence are used to form a Markov transition kernel on the space of hypotheses. The approach leads to easy interpretable results and involves very simple formulas suitable to analyze large datasets as those arising from gene expression data (microarray or RNA-seq experiments).

Correction: Energy expenditure in caving.

[This corrects the article DOI: 10.1371/journal.pone.0170853.].

Helicobacter pylori and intestinal parasites are not detrimental to the nutritional status of Amerindians.

Gastrointestinal parasites have evolved with humans and colonize many asymptomatic subjects. We investigated the influence of microbial gastrointestinal colonization on the nutritional status of rural Amerindians (40 males and 61 females). Helicobacter pylori was detected by 13C-breath test, and intestinal parasites were detected in fecal specimens. Body morphometry and bioelectrical impedance measurements were measured. Although Amerindians showed low height and weight for age, they had an adequate body mass index, morphometric parameters, and cell mass. Intestinal parasites were detected in 99% of the subjects, with no detrimental effect on nutritional parameters. Helicobacter pylori was present in 82% of adults and half the children, and was positively correlated with improved nutritional status. Despite the high prevalence of gastrointestinal microbes often associated with disease, the studied population of Amerindians had a body morphometry and composition indicative of good nutritional status, and improved in children positive for gastric H. pylori.

Patterns of hand variation--new data on a Sardinian sample.

This study is an analysis of the patterns of variation of the human hand, particularly the metric characters of palm, fingers and distal phalanges. Anthropometric measurements were performed on 146 Sardinian men and women, aged 21 to 31 years. The data were analyzed by inferential statistics (paired Student's t test, analysis of variance), and Principal Components Analysis. The results indicate that size factors are the principal source of variation. A residual adimensional component of variability is related to diversification between the fingers as a whole and the distal phalanges, and between the thumb and the other fingers. Sexual dimorphism is evident. Men present greater dimensions and greater relative length of the thumb with respect to the other fingers than women.

The taste of heat: How humoral qualities act as a cultural filter for chemosensory properties guiding herbal medicine.

ETHNOPHARMACOLOGICAL RELEVANCE: Organoleptic properties, and more specifically chemosensory cues, have been shown to guide therapeutic applications of medicinal plants. Humoral qualities, on the other hand, are widely believed to be an abstract concept, mainly applied post hoc to validate therapy. However, the nexus between humoral qualities, chemosensory properties, and medicinal plant uses has never been systematically assessed. AIM OF THE STUDY: To systematically analyse the correlations between chemosensory properties, humoral qualities, and medicinal uses of selected botanical drugs. METHODS: The issue was approached experimentally via an organoleptic testing panel, consisting of Zoque healers in Chiapas, Mexico. The healers smelled and tasted 71 selected herbal drugs and subsequently commented on their humoral qualities and therapeutic uses. The resulting dataset is analysed for correlations between these variables using Bayesian statistics. Qualitative data on the characteristics and role of the hot-cold dichotomy complement the quantitative analysis, facilitating meaningful interpretation. RESULTS AND DISCUSSION: The results reproduce and extend the findings of previous studies, which established specific correlations between chemosensory cues and nosological units. The key predictors of drugs' therapeutic uses, however, are their humoral qualities, which are themselves conditioned by taste and smell. These findings appear to be valid for drug samples known to the participants as well as for unfamiliar samples. Thus, this study establishes the role of the hot-cold dichotomy as an important cultural filter connecting organoleptic properties and therapeutic uses of herbal drugs. CONCLUSIONS: There is considerable cross-cultural consensus in Mesoamerica for the specific correlations described in this study. Given the continued pervasiveness of the hot-cold dichotomy, humoral qualities and the underlying organoleptic properties ought to be increasingly considered in the design of pharmaceutical products as well as public health strategies. Such culturally appropriate adjustments may considerably improve the perceived quality and effectiveness of healthcare.