Bayesian longitudinal data analysis with mixed models and thick-tailed distributions using MCMC

Linear mixed effects models are frequently used to analyse longitudinal data, due to their flexibility in modelling the covariance structure between and within observations. Further, it is easy to deal with unbalanced data, either with respect to the number of observations per subject or per time period, and with varying time intervals between observations. In most applications of mixed models to biological sciences, a normal distribution is assumed both for the random effects and for the residuals. This, however, makes inferences vulnerable to the presence of outliers. Here, linear mixed models employing thick-tailed distributions for robust inferences in longitudinal data analysis are described. Specific distributions discussed include the Student-t, the slash and the contaminated normal. A Bayesian framework is adopted, and the Gibbs sampler and the Metropolis-Hastings algorithms are used to carry out the posterior analyses. An example with data on orthodontic distance growth in children is discussed to illustrate the methodology. Analyses based on either the Student-t distribution or on the usual Gaussian assumption are contrasted. The thick-tailed distributions provide an appealing robust alternative to the Gaussian process for modelling distributions of the random effects and of residuals in linear mixed models, and the MCMC implementation allows the computations to be performed in a flexible manner.

Quantum and semiclassical Husimi distributions for a one-dimensional resonant system

We compare exact and semiclassical Husimi distributions for the single eigenstates of a one-dimensional resonant Hamiltonian. We find that both distributions concentrate near the unstable fixed points even when these points are made complex by suitably varying a parameter. © 1992 The American Physical Society.

Robust linear mixed models with normal/independent distributions and Bayesian MCMC implementation

Linear mixed effects models have been widely used in analysis of data where responses are clustered around some random effects, so it is not reasonable to assume independence between observations in the same cluster. In most biological applications, it is assumed that the distributions of the random effects and of the residuals are Gaussian. This makes inferences vulnerable to the presence of outliers. Here, linear mixed effects models with normal/independent residual distributions for robust inferences are described. Specific distributions examined include univariate and multivariate versions of the Student-t, the slash and the contaminated normal. A Bayesian framework is adopted and Markov chain Monte Carlo is used to carry out the posterior analysis. The procedures are illustrated using birth weight data on rats in a texicological experiment. Results from the Gaussian and robust models are contrasted, and it is shown how the implementation can be used for outlier detection. The thick-tailed distributions provide an appealing robust alternative to the Gaussian process in linear mixed models, and they are easily implemented using data augmentation and MCMC techniques.

Allele frequency distributions of six hypervariable loci (D1S80, APOB, D4S43, vW1, F13A and DYS19) in two African-Brazilian communities from the Amazon region

ABSTRACT: The allele frequency distributions of three VNTR (D1S80, APOB and D4S43) and three STR (vW1, F13A1 and DYS19) loci were investigated in two Afro-Brazilian populations from the Amazon: Curiau and Pacoval. Exact tests for population differentiation revealed significant differences in allele frequency between populations only for the D1S80 and APOB loci. A statistically significant deviation from the Hardy-Weinberg equilibrium was observed only in the D1S80 locus of the Pacoval sample. A neighbor-joining tree was constructed based on DA genetic distances of allele frequencies in four Afro-Brazilian populations from the Amazon (Pacoval, Curiau, Trombetas, and Cametá), along with those from Congo, Cameroon, Brazilian Amerindians, and Europeans. This analysis revealed the usefulness of these Amp-FLPs for population studies - African and African-derived populations were closely grouped, and clearly separated from Amerindians and Europeans. Estimates of admixture components based on the gene identity method revealed the prevalence of the African component in both populations studied, amounting to 51% in Pacoval, and to 43% in Curiau. The Amerindian component was also important in both populations (37% in Pacoval, and 24% in Curiau). The European component reached 33% in Curiau.

Distributions and phylogeographic data of rheophilic freshwater fishes provide evidences on the geographic extension of a central-brazilian amazonian palaeoplateau in the area of the present day Pantanal Wetland

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Rapidity distributions in exclusive Z plus jet and gamma plus jet events in pp collisions at root s=7 TeV

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Single-Particle Momentum Distributions for Bosonic Trimer States in Two and Three Dimensions

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Linkage disequilibrium levels in Bos indicus and Bos taurus cattle using medium and high density SNP chip data and different minor allele frequency distributions

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Search for quark contact interactions and extra spatial dimensions using dijet angular distributions in proton–proton collisions at √s = 8 TeV

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Measurement of pseudorapidity distributions of charged particles in proton-proton collisions at root s=8 TeV by the CMS and TOTEM experiments

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Drosophila melanogaster lipins are tissue-regulated and developmentally regulated and present specific subcellular distributions

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Power laws, Discontinuities and Regional City Size Distributions

Urban systems are manifestations of human adaptation to the natural environment. City size distributions are the expression of hierarchical processes acting upon urban systems. In this paper, we test the entire city size distributions for the southeastern and southwestern United States (1990), as well as the size classes in these regions for power law behavior. We interpret the differences in the size of the regional city size distributions as the manifestation of variable growth dynamics dependent upon city size. Size classics in the city size distributions are snapshots of stable states within urban systems in flux.

Environmental suitability of a highly fragmented and heterogeneous landscape for forest bird species in south-eastern Brazil

Assessment of the suitability of anthropogenic landscapes for wildlife species is crucial for setting priorities for biodiversity conservation. This study aimed to analyse the environmental suitability of a highly fragmented region of the Brazilian Atlantic Forest, one of the world's 25 recognized biodiversity hotspots, for forest bird species. Eight forest bird species were selected for the analyses, based on point counts (n = 122) conducted in April-September 2006 and January-March 2009. Six additional variables (landscape diversity, distance from forest and streams, aspect, elevation and slope) were modelled in Maxent for (1) actual and (2) simulated land cover, based on the forest expansion required by existing Brazilian forest legislation. Models were evaluated by bootstrap or jackknife methods and their performance was assessed by AUC, omission error, binomial probability or p value. All predictive models were statistically significant, with high AUC values and low omission errors. A small proportion of the actual landscape (24.41 +/- 6.31%) was suitable for forest bird species. The simulated landscapes lead to an increase of c. 30% in total suitable areas. In average, models predicted a small increase (23.69 +/- 6.95%) in the area of suitable native forest for bird species. Being close to forest increased the environmental suitability of landscapes for all bird species; landscape diversity was also a significant factor for some species. In conclusion, this study demonstrates that species distribution modelling (SDM) successfully predicted bird distribution across a heterogeneous landscape at fine spatial resolution, as all models were biologically relevant and statistically significant. The use of landscape variables as predictors contributed significantly to the results, particularly for species distributions over small extents and at fine scales. This is the first study to evaluate the environmental suitability of the remaining Brazilian Atlantic Forest for bird species in an agricultural landscape, and provides important additional data for regional environmental planning.

Bivariate gamma-geometric law and its induced Levy process

In this article we introduce a three-parameter extension of the bivariate exponential-geometric (BEG) law (Kozubowski and Panorska, 2005) [4]. We refer to this new distribution as the bivariate gamma-geometric (BGG) law. A bivariate random vector (X, N) follows the BGG law if N has geometric distribution and X may be represented (in law) as a sum of N independent and identically distributed gamma variables, where these variables are independent of N. Statistical properties such as moment generation and characteristic functions, moments and a variance-covariance matrix are provided. The marginal and conditional laws are also studied. We show that BBG distribution is infinitely divisible, just as the BEG model is. Further, we provide alternative representations for the BGG distribution and show that it enjoys a geometric stability property. Maximum likelihood estimation and inference are discussed and a reparametrization is proposed in order to obtain orthogonality of the parameters. We present an application to a real data set where our model provides a better fit than the BEG model. Our bivariate distribution induces a bivariate Levy process with correlated gamma and negative binomial processes, which extends the bivariate Levy motion proposed by Kozubowski et al. (2008) [6]. The marginals of our Levy motion are a mixture of gamma and negative binomial processes and we named it BMixGNB motion. Basic properties such as stochastic self-similarity and the covariance matrix of the process are presented. The bivariate distribution at fixed time of our BMixGNB process is also studied and some results are derived, including a discussion about maximum likelihood estimation and inference. (C) 2012 Elsevier Inc. All rights reserved.