Search

Bayesian methods provide direct uncertainty quantification in functional data analysis applications without reliance on bootstrap techniques. A major tool in functional data applications is the functional principal component analysis which decomposes the data around a common mean function and identifies leading directions of variation. Bayesian functional principal components analysis (BFPCA) provides uncertainty quantification on the estimated functional model components via the posterior samples obtained. We propose central posterior envelopes (CPEs) for BFPCA based on functional depth as a descriptive visualization tool to summarize variation in the posterior samples of the estimated functional model components, contributing to uncertainty quantification in BFPCA. The proposed BFPCA relies on a latent factor model and targets model parameters within a hierarchical modeling framework using modified multiplicative gamma process shrinkage priors on the variance components. Functional depth provides a center-outward order to a sample of functions. We utilize modified band depth and modified volume depth for ordering of a sample of functions and surfaces, respectively, to derive at CPEs of the mean and eigenfunctions within the BFPCA framework. The proposed CPEs are showcased in extensive simulations. Finally, the proposed CPEs are applied to the analysis of a sample of power spectral densities from resting state electroencephalography where they lead to novel insights on diagnostic group differences among children diagnosed with autism spectrum disorder and their typically developing peers across age.

Background: Weight gain, insulin-like growth factor-I (IGF-I) levels, and excess exogenous steroid hormone use are putative cancer risk factors, yet their interconnected pathways have not been fully characterized. This cross-sectional study investigated the relationship between plasma IGF-I levels and weight gain according to body mass index (BMI), leptin levels, and exogenous estrogen use among postmenopausal women.Methods: This study included 794 postmenopausal women who enrolled in an ancillary study of the Women's Health Initiative Observational Study between February 1995 and July 1998. The relationship between IGF-I levels and weight gain was analyzed using ordinal logistic regression. We used the molar ratio of IGF-I to IGF binding protein-3 (IGF-I/IGFBP-3) or circulating IGF-I levels adjusting for IGFBP-3 as a proxy of bioavailable IGF-I. The plasma concentrations were expressed as quartiles.Results: Among the obese group, women in the third quartile (Q3) of IGF-I and highest quartile of IGF-I/IGFBP-3 were less likely to gain weight (>3% from baseline) than were women in the first quartiles (Q1). Among the normal-weight group, women in Q2 and Q3 of IGF-I/IGFBP-3 were 70% less likely than those in Q1 to gain weight. Among current estrogen users, Q3 of IGF-I/IGFBP-3 had 0.5 times the odds of gaining weight than Q1.Conclusions: Bioavailable IGF-I levels were inversely related to weight gain overall.Impact: Although weight gain was not consistent with increases in IGF-I levels among postmenopausal women in this report, avoidance of weight gain as a strategy to reduce cancer risk may be recommended. Cancer Epidemiol Biomarkers Prev; 23(3); 534–44. ©2013 AACR.

PDF file - 70 KB, Supplemental Table S1. Median IGF concentrations by exogenous estrogen use according to weight-change group. Supplemental Table S2. Distributions of Insulin by weight change group. Supplemental Figure S1. Median IGF concentrations by weight change group according to body mass index.

The analysis of large-scale datasets, especially in biomedical contexts, frequently involves a principled screening of multiple hypotheses. The celebrated two-group model jointly models the distribution of the test statistics with mixtures of two competing densities, the null and the alternative distributions. We investigate the use of weighted densities and, in particular, non-local densities as working alternative distributions, to enforce separation from the null and thus refine the screening procedure. We show how these weighted alternatives improve various operating characteristics, such as the Bayesian False Discovery rate, of the resulting tests for a fixed mixture proportion with respect to a local, unweighted likelihood approach. Parametric and nonparametric model specifications are proposed, along with efficient samplers for posterior inference. By means of a simulation study, we exhibit how our model compares with both well-established and state-of-the-art alternatives in terms of various operating characteristics. Finally, to illustrate the versatility of our method, we conduct three differential expression analyses with publicly-available datasets from genomic studies of heterogeneous nature.

INTRODUCTION Although humans are capable of enduring critically low levels of oxygen, many hypoxaemic patients die despite aggressive therapies. Mimicking the physiological hyperventilation necessary to survive extreme hypoxic conditions could minimize the derangements caused by acute hypoxic-hypoxia. The objective of this study was to measure the haemodynamic-biochemical response to artificially induced hyperventilation in hypoxic rats. MATERIAL AND METHODS Twenty-four deeply anaesthetized and mechanically ventilated rats were allocated to 3 groups: control (n = 5, FiO2 = 1); hypoxic spontaneously hyperventilating (n = 10, FiO2 = 0.08); and hypoxic artificially induced hyperventilation (n = 9, targeting PaCO2 = 10 mm Hg, FiO2 = 0.08). We compared the spontaneously and artificially hyperventilating groups. P-values < 0.01 were considered statistically significant. Mean arterial pressure (MAP) and serum chemistry were measured for 180 minutes. RESULTS The control group remained stable throughout the experiment. The hypoxic groups developed profound hypotension after the decrease in FiO2. However, the artificially induced hyperventilated rats recovered their MAP to levels higher than the spontaneously hyperventilating group (117.1 ± 17.2 vs. 68.1 ± 16.0, P = 0.0048). In regard to the biochemical derangements, even though the serum lactate and PaO2 were not different among the hypoxic groups, the artificially hyperventilated group achieved significantly higher SaO2 (94.3 ± 3.6 vs. 58.6 ± 9.6, P = 0.005), pH (7.87 ± 0.04 vs. 7.50 ± 0.13, P = 0.005), and CaO2 (17.7 ± 2.6 vs. 10.2 ± 1.3, P = 0.005) at 180 minutes. CONCLUSIONS Artificially induced hyperventilation led to the correction of arterial oxygen content, severe serum chemistry, and haemodynamic derangements. These findings may represent a novel rescue manoeuvre and serve as a bridge to a permanent form of support, but should be further studied before being translated to the clinical setting.

Background: Three national career development programs (CDPs)—Early and Mid-Career Programs sponsored by the Association of American Medical Colleges and the Hedwig van Ameringen Executive Leadership in Academic Medicine sponsored by Drexel University—seek to expand gender diversity in faculty and institutional leadership of academic medical centers. Over 20 years of success and continued need are evident in the sustained interest and investment of individuals and institutions. However, their impact on promotion in academic rank remains unknown. The purpose of the study is to compare promotion rates of women CDP participants and other faculty of similar institutional environment and initial career stage. Methods: The study examined retrospective cohorts of 2,719 CDP participants, 12,865 nonparticipant women, and 26,810 men, from the same institutions, with the same degrees, and first years of appointment in rank. Rates of promotion to Associate and Full Professor ranks in respective cohorts of Assistant and of Associate Professors were compared using Kaplan-Meier survival curves and log-rank tests, and logistic regression adjusting for other predictors of academic success. Results: In adjusted analyses, participants were more likely than men and non-participant women to be promoted to Associate Professor and as likely as men and more likely than non-participant women to be promoted to Full Professor within 10 years. Within 5 years, CDP participants were more likely than nonparticipant women to be promoted to Associate Professor and as likely as to be promoted to Full Professor; in the same interval, participants were promoted to both higher ranks at the same rates as men. For both intervals, nonparticipant women were significantly less likely than men to be promoted to either rank. Conclusions: The higher rates of promotion for women participating in national CDPs support the effectiveness of these programs in building capacity for academic medicine.

Partially exchangeable datasets are characterized by observations grouped into known, heterogeneous units. The recently developed Common Atoms Model (CAM) is a Bayesian nonparametric technique suited for analyzing this type of data. CAM induces a two-layered clustering structure: one across observations and another across units. In particular, the units are clustered according to their distributional similarities. In this article, we illustrate the versatility of CAM with an application to an openly available Spotify dataset. The dataset contains quantitative audio features for a large number of songs grouped by artists. After describing the data preprocessing steps, we employ CAM to group the Spotify artists according to the distributions of the energy of their songs.

Summary In basic neuroscience research, data are often clustered or collected with repeated measures, hence correlated. The most widely used methods such as t test and ANOVA do not take data dependence into account and thus are often misused. This Primer introduces linear and generalized mixed-effects models that consider data dependence and provides clear instruction on how to recognize when they are needed and how to apply them. The appropriate use of mixed-effects models will help researchers improve their experimental design and will lead to data analyses with greater validity and higher reproducibility of the experimental findings.

Human platelet lysate (hPL) is considered a valid substitute to fetal bovine serum (FBS) in the expansion of mesenchymal stromal cells (MSC), and it is commonly produced starting from intermediate side products of whole blood donations. Through freeze–thaw cycles, hPL is highly enriched in chemokines, growth factors, and adhesion and immunologic molecules. Cell therapy protocols, using hPL instead of FBS for the expansion of cells, are approved by regulatory authorities without concerns, and its administration in patients is considered safe. However, published data are fairly difficult to compare, since the production of hPL is highly variable. This study proposes to optimize and standardize the hPL productive process by using instruments, technologies, and quality/safety standards required for blood bank activities and products. The quality and improved selection of the starting material (i.e., the whole blood), together with the improvement of the production process, guarantee a product characterized by higher content and quality of growth factors as well as a reduction in batch-to-batch variability. By increasing the number of freeze/thaw cycles from one (hPL1c) to four (hPL4c), we obtained a favorable effect on the release of growth factors from platelet α granules. Those changes have directly translated into biological effects leading to a decreasing doubling time (DT) of MSC expansion at 7 days (49.41 ± 2.62 vs. 40.61 ± 1.11 h, p < 0.001). Furthermore, mass spectrometry (MS)-based evaluation has shown that the proliferative effects of hPL4c are also combined with a lower batch-to-batch variability (10–15 vs. 21–31%) at the proteomic level. In conclusion, we have considered lot-to-lot hPL variability, and by the strict application of blood bank standards, we have obtained a standardized, reproducible, safe, cheap, and ready-to-use product.

In this article, we develop a Bayesian hierarchical mixture regression model for studying the association between a multivariate response, measured as counts on a set of features, and a set of covariates. We have available RNA-Seq and DNA methylation data measured on breast cancer patients at different stages of the disease. We account for the heterogeneity and over-dispersion of count data (here, RNA-Seq data) by considering a mixture of negative binomial distributions and incorporate the covariates (here, methylation data) into the model via a linear modeling construction on the mean components. Our modeling construction includes several innovative characteristics. First, it employs selection techniques that allow the identification of a small subset of features that best discriminate the samples while simultaneously selecting a set of covariates associated to each feature. Second, it incorporates known dependencies into the feature selection process via the use of Markov random field (MRF) priors. On simulated data, we show how incorporating existing information via the prior model can improve the accuracy of feature selection. In the analysis of RNA-Seq and DNA methylation data on breast cancer, we incorporate knowledge on relationships among genes via a gene-gene network, which we extract from the KEGG database. Our data analysis identifies genes which are discriminatory of cancer stages and simultaneously selects significant associations between those genes and DNA methylation sites. A biological interpretation of our findings reveals several biomarkers that can help understanding the effect of DNA methylation on gene expression transcription across cancer stages.

The use of large datasets for targeted therapeutic interventions requires new ways to characterize the heterogeneity observed across subgroups of a specific population. In particular, models for partially exchangeable data are needed for inference on nested datasets, where the observations are assumed to be organized in different units and some sharing of information is required to learn distinctive features of the units. In this manuscript, we propose a nested common atoms model (CAM) that is particularly suited for the analysis of nested datasets where the distributions of the units are expected to differ only over a small fraction of the observations sampled from each unit. The proposed CAM allows a two-layered clustering at the distributional and observational level and is amenable to scalable posterior inference through the use of a computationally efficient nested slice sampler algorithm. We further discuss how to extend the proposed modeling framework to handle discrete measurements, and we conduct posterior inference on a real microbiome dataset from a diet swap study to investigate how the alterations in intestinal microbiota composition are associated with different eating habits. We further investigate the performance of our model in capturing true distributional structures in the population by means of a simulation study.

The simultaneous testing of multiple hypotheses is common to the analysis of high-dimensional data sets. The two-group model, first proposed by Efron, identifies significant comparisons by allocating observations to a mixture of an empirical null and an alternative distribution. In the Bayesian nonparametrics literature, many approaches have suggested using mixtures of Dirichlet Processes in the two-group model framework. Here, we investigate employing mixtures of two-parameter Poisson-Dirichlet Processes instead, and show how they provide a more flexible and effective tool for large-scale hypothesis testing. Our model further employs nonlocal prior densities to allow separation between the two mixture components. We obtain a closed-form expression for the exchangeable partition probability function of the two-group model, which leads to a straightforward Markov Chain Monte Carlo implementation. We compare the performance of our method for large-scale inference in a simulation study and illustrate its use on both a prostate cancer data set and a case-control microbiome study of the gastrointestinal tracts in children from underdeveloped countries who have been recently diagnosed with moderate-to-severediarrhea.

Electroencephalography (EEG) is a noninvasive neuroimaging modality that captures electrical brain activity many times per second. We seek to estimate power spectra from EEG data that ware gathered for 557 adolescent twin pairs through the Minnesota Twin Family Study (MTFS). Typically, spectral analysis methods treat time series from each subject separately, and independent spectral densities are fit to each time series. Since the EEG data were collected on twins, it is reasonable to assume that the time series have similar underlying characteristics, so borrowing information across subjects can significantly improve estimation. We propose a Nested Bernstein Dirichlet prior model to estimate the power spectrum of the EEG signal for each subject by smoothing periodograms within and across subjects while requiring minimal user input to tuning parameters. Furthermore, we leverage the MTFS twin study design to estimate the heritability of EEG power spectra with the hopes of establishing new endophenotypes. Through simulation studies designed to mimic the MTFS, we show our method out-performs a set of other popular methods.

To make predictions about the carbon cycling consequences of rising global surface temperatures, Earth system scientists rely on mathematical soil biogeochemical models (SBMs). However, it is not clear which models have better predictive accuracy, and a rigorous quantitative approach for comparing and validating the predictions has yet to be established. In this study, we present a Bayesian approach to SBM comparison that can be incorporated into a statistical model selection framework. We compared the fits of linear and nonlinear SBMs to soil respiration data compiled in a recent meta-analysis of soil warming field experiments. Fit quality was quantified using Bayesian goodness-of-fit metrics, including the widely applicable information criterion (WAIC) and leave-one-out cross validation (LOO). We found that the linear model generally outperformed the nonlinear model at fitting the meta-analysis data set. Both WAIC and LOO computed higher overfitting risk and effective numbers of parameters for the nonlinear model compared to the linear model, conditional on the data set. Goodness of fit for both models generally improved when they were initialized with lower and more realistic steady-state soil organic carbon densities. Still, testing whether linear models offer definitively superior predictive performance over nonlinear models on a global scale will require comparisons with additional site-specific data sets of suitable size and dimensionality. Such comparisons can build upon the approach defined in this study to make more rigorous statistical determinations about model accuracy while leveraging emerging data sets, such as those from long-term ecological research experiments.

There is a strong interest in the neuroscience community to measure brain connectivity and develop methods that can differentiate connectivity across patient groups and across different experimental stimuli. The development of such statistical tools is critical to understand the dynamics of functional relationships among brain structures supporting memory encoding and retrieval. However, challenges arise by providing from the need to incorporate within-condition similarity with between-conditions heterogeneity in modeling connectivity, as well as how to provide a natural way to conduct trial- and condition-level inference on effective connectivity. A Bayesian hierarchical vector autoregressive (BH-VAR) model is proposed to characterize brain connectivity and infer differences in connectivity across conditions. Within-condition connectivity similarity and between-conditions connectivity heterogeneity are accounted for by the priors on trial-specific models. In addition to the fully Bayesian framework, an alternative two-stage computational approach is also proposed which still allows straightforward uncertainty quantification of between-trial conditions via MCMC posterior sampling, but provides a fast approximate procedure for the estimation of trial-specific VAR parameters. A novel aspect of the approach is the use of a frequency-specific measure, partial directed coherence (PDC), to characterize effective connectivity under the Bayesian framework. More specifically, PDC allows inferring directionality and explaining the extent to which the present oscillatory activity at a certain frequency in a sender channel influences the future oscillatory activity in a specific receiver channel relative to all possible receivers in the brain network. The proposed model is applied to a large electrophysiological dataset collected as rats performed a complex sequence memory task. This unique dataset includes local field potentials (LFPs) activity recorded from an array of electrodes across the hippocampal region CA1 while animals were presented with multiple trials from two main conditions. The proposed modeling approach provided novel insights into hippocampal connectivity during memory performance. Specifically, it separated CA1 into two functional units, a lateral and a medial segment, each showing stronger functional connectivity to itself than to the other. This approach also revealed that information primarily flowed in a lateral-to-medial direction across trials (within-condition), and suggested this effect was stronger on one trial condition than the other (between-conditions effect). Collectively, these results indicate that the proposed model is a promising approach to quantify the variation of functional connectivity, both within- and between-conditions, and thus should have broad applications in neuroscience research.

Atherosclerosis represents one of the main risk factors for the development of cardiovascular diseases. Their etiologies have been studied in recent years in order to better define therapeutic targets for intervention and to identify diagnostic methods. Two different subtypes of macrophages, M1 and M2, have been described in physiological conditions. They can also be found in the atherosclerotic process, where they both have opposite roles in disease progression. Perivascular brown adipose tissue is also involved in inflammation and endothelial damage. In this work, we provide insights into the protective role of melatonin in the atherosclerotic process by morphological and 18F-FDG-PET/CT analyses. In particular, we examined the effects of melatonin on pathways that are linked to atherosclerosis development. We showed that melatonin, by suppressing M1 activity, reduced inflammation and directed macrophage polarization toward the M2 macrophage subtype. Moreover, melatonin preserved the activity of perivascular brown adipose tissue. In addition, 18F-FDG uptake is very high in mice treated with melatonin, confirming that other factors may alter 18F-FDG distribution. In conclusion, we showed that melatonin affects inflammatory pathways that have been linked to atherosclerosis, assessed the relationships of the 18F-FDG PET/CT parameters with macrophage markers and the production of their cytokines, which that have been defined by morphological evaluations.

Cancer radiomics is an emerging discipline promising to elucidate lesion phenotypes and tumor heterogeneity through patterns of enhancement, texture, morphology, and shape. The prevailing technique for image texture analysis relies on the construction and synthesis of Gray-Level Co-occurrence Matrices (GLCM). Practice currently reduces the structured count data of a GLCM to reductive and redundant summary statistics for which analysis requires variable selection and multiple comparisons for each application, thus limiting reproducibility. In this article, we develop a Bayesian multivariate probabilistic framework for the analysis and unsupervised clustering of a sample of GLCM objects. By appropriately accounting for skewness and zero-inflation of the observed counts and simultaneously adjusting for existing spatial autocorrelation at nearby cells, the methodology facilitates estimation of texture pattern distributions within the GLCM lattice itself. The techniques are applied to cluster images of adrenal lesions obtained from CT scans with and without administration of contrast. We further assess whether the resultant subtypes are clinically oriented by investigating their correspondence with pathological diagnoses. Additionally, we compare performance to a class of machine-learning approaches currently used in cancer radiomics with simulation studies., Comment: 45 pages, 7 figures, 1 Table

Variable selection, also known as feature selection in the machine learning literature, plays an indispensable role in scientific studies. In many research areas with massive data, finding a subset of representative features that best explain the outcome of interest has become a critical component in any researcher's workflow. In this chapter, we focus on Bayesian variable selection regression models for count data, and specifically on the negative binomial linear regression model and on the Dirichlet-multinomial regression model. We address the variable selection problem via spike-and-slab priors. For posterior inference, we review standard MCMC methods and also investigate computationally more efficient variational inference approaches that use data augmentation techniques and concrete relaxation methods. We investigate performance of the methods via simulation studies and benchmark data sets.

The objectives of the study are (1) to distinguish lipoma (L) from atypical lipomatous tumor (ALT) using MRI qualitative features, (2) to assess the value of contrast enhancement, and (3) to evaluate the reproducibility and confidence level of radiological readings. Patients with pathologically proven L or ALT, who underwent MRI within 3 months from surgical excision were included in this retrospective multicenter international study. Two radiologists independently reviewed MRI centrally. Impressions were recorded as L or ALT. A third radiologist was consulted for discordant readings. The two radiologists re-read all non-contrast sequences; impression was recorded; then post-contrast images were reviewed and any changes were recorded. A total of 246 patients (135 females; median age, 59 years) were included. ALT was histopathologically confirmed in 70/246 patients. In multivariable analysis, in addition to the lesion size, deep location, proximal lower limb lesions, demonstrating incomplete fat suppression, or increased architectural complexity were the independent predictive features of ALT; but not the contrast enhancement. Post-contrast MRI changed the impression in a total of 5 studies (3 for R1 and 4 for R2; 2 studies are common); all of them were incorrectly changed from Ls to ALTs. Overall, inter-reader kappa agreement was 0.42 (95% CI 0.39–0.56). Discordance between the two readers was statistically significant for both pathologically proven L (p