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.

Generalized linear mixed models for the genetic evaluation of binary reproductive traits: a simulation study

The objective of this study was to evaluate the use of probit and logit link functions for the genetic evaluation of early pregnancy using simulated data. The following simulation/analysis structures were constructed: logit/logit, logit/probit, probit/logit, and probit/probit. The percentages of precocious females were 5, 10, 15, 20, 25 and 30% and were adjusted based on a change in the mean of the latent variable. The parametric heritability (h²) was 0.40. Simulation and genetic evaluation were implemented in the R software. Heritability estimates (ĥ²) were compared with h² using the mean squared error. Pearson correlations between predicted and true breeding values and the percentage of coincidence between true and predicted ranking, considering the 10% of bulls with the highest breeding values (TOP10) were calculated. The mean ĥ² values were under- and overestimated for all percentages of precocious females when logit/probit and probit/logit models used. In addition, the mean squared errors of these models were high when compared with those obtained with the probit/probit and logit/logit models. Considering ĥ², probit/probit and logit/logit were also superior to logit/probit and probit/logit, providing values close to the parametric heritability. Logit/probit and probit/logit presented low Pearson correlations, whereas the correlations obtained with probit/probit and logit/logit ranged from moderate to high. With respect to the TOP10 bulls, logit/probit and probit/logit presented much lower percentages than probit/probit and logit/logit. The genetic parameter estimates and predictions of breeding values of the animals obtained with the logit/logit and probit/probit models were similar. In contrast, the results obtained with probit/logit and logit/probit were not satisfactory. There is need to compare the estimation and prediction ability of logit and probit link functions.

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.

A class of mixed integrable models

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Linear and Poisson models for genetic evaluation of tick resistance in cross-bred Hereford x Nellore cattle

Cattle resistance to ticks is measured by the number of ticks infesting the animal. The model used for the genetic analysis of cattle resistance to ticks frequently requires logarithmic transformation of the observations. The objective of this study was to evaluate the predictive ability and goodness of fit of different models for the analysis of this trait in cross-bred Hereford x Nellore cattle. Three models were tested: a linear model using logarithmic transformation of the observations (MLOG); a linear model without transformation of the observations (MLIN); and a generalized linear Poisson model with residual term (MPOI). All models included the classificatory effects of contemporary group and genetic group and the covariates age of animal at the time of recording and individual heterozygosis, as well as additive genetic effects as random effects. Heritability estimates were 0.08 ± 0.02, 0.10 ± 0.02 and 0.14 ± 0.04 for MLIN, MLOG and MPOI models, respectively. The model fit quality, verified by deviance information criterion (DIC) and residual mean square, indicated fit superiority of MPOI model. The predictive ability of the models was compared by validation test in independent sample. The MPOI model was slightly superior in terms of goodness of fit and predictive ability, whereas the correlations between observed and predicted tick counts were practically the same for all models. A higher rank correlation between breeding values was observed between models MLOG and MPOI. Poisson model can be used for the selection of tick-resistant animals. © 2013 Blackwell Verlag GmbH.

Genetic analysis of the temperament of Nellore cattle using linear and threshold models

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Community-wide spatial and temporal discordances of seed-seedling shadows in a tropical rainforest

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

Estudo de simulação de modelos lineares mistos com distribuição normal contaminada no melhoramento genético animal

Objetivou-se com esse trabalho comparar estimativas de componentes de variâncias obtidas por meio de modelos lineares mistos Gaussianos e Robustos, via Amostrador de Gibbs, em dados simulados. Foram simulados 50 arquivos de dados com 1.000 animais cada um, distribuídos em cinco gerações, em dois níveis de efeito fixo e três valores fenotípicos distintos para uma característica hipotética, com diferentes níveis de contaminação. Exceto para os dados sem contaminação, quando os modelos foram iguais, o modelo Robusto apresentou melhores estimativas da variância residual. As estimativas de herdabilidade foram semelhantes em todos os modelos, mas as análises de regressão mostraram que os valores genéticos preditos com uso do modelo Robusto foram mais próximos dos valores genéticos verdadeiros. Esses resultados sugerem que o modelo linear normal contaminado oferece uma alternativa flexível para estimação robusta em melhoramento genético animal.

Genetic and non-genetic effects on calf vigor at birth and preweaning mortality in nellore calves

The objective of the present study was to evaluate the genetic and non-genetic effects that influencevigor at birth and preweaning mortality in Nellore calves. A total of 11,727 records of births that occurred between 1978 and 2006, offspring of 363 sires, were analyzed. Poor calf vigor at birth (VB) and preweaning mortality divided into stillbirth (SB), early mortality (EM) and total mortality (TM) were analyzed as binary variables. Generalized linear models were used for the evaluation of non-genetic effects and generalized linear mixed models for genetic effects (sire and animal models). The incidences were 4.75% for VB, 2.66% for SB, 5.28% for EM, and 7.99% for TM. Birth weight was the effect that most influenced the traits studied. Calves weighing less than 22kg(females) and less than 24kg (males) were at a higher risk of low vigor and preweaning mortality. Preweaning mortality was higher among calves born from cows aged .3 and .11 years at calving compared with cows aged 7 to 10 years. Male calves presented less vigor and higher preweaning mortality than female calves. Selection for postweaning weight did not influence preweaning mortality. The heritability estimates ranged between 0.01 and 0.09 for VB, 0.00 and 0.27 for SB, 0.03 and 0.17 for EM and 0.02 and 0.10 for TM. Stillbirth should be included as a selection criterion in breeding programs of Nellore cattle, alone or as part of a selection index, aiming to reduce preweaning mortality. © 2013 Sociedade Brasileira de Zootecnia.

Estimativas de parâmetros genéticos para escores visuais e características de crescimento na raça Nelore

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Variabilidades genética e não genética na mortalidade pré-desmama de bezerros de corte

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Allometric equations for estimating tree biomass in restored mixed-species Atlantic Forest stands

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

Análise genética de escores de avaliação visual de bovinos com modelos bayesianos de limiar e linear

O objetivo deste trabalho foi comparar as estimativas de parâmetros genéticos obtidas em análises bayesianas uni-característica e bi-característica, em modelo animal linear e de limiar, considerando-se as características categóricas morfológicas de bovinos da raça Nelore. Os dados de musculosidade, estrutura física e conformação foram obtidos entre 2000 e 2005, em 3.864 animais de 13 fazendas participantes do Programa Nelore Brasil. Foram realizadas análises bayesianas uni e bi-características, em modelos de limiar e linear. de modo geral, os modelos de limiar e linear foram eficientes na estimação dos parâmetros genéticos para escores visuais em análises bayesianas uni-características. Nas análises bi-características, observou-se que: com utilização de dados contínuos e categóricos, o modelo de limiar proporcionou estimativas de correlação genética de maior magnitude do que aquelas do modelo linear; e com o uso de dados categóricos, as estimativas de herdabilidade foram semelhantes. A vantagem do modelo linear foi o menor tempo gasto no processamento das análises. Na avaliação genética de animais para escores visuais, o uso do modelo de limiar ou linear não influenciou a classificação dos animais, quanto aos valores genéticos preditos, o que indica que ambos os modelos podem ser utilizados em programas de melhoramento genético.

Analysis of the electrical conductivity in milking fractions as a mean for detecting and characterizing mastitis in goats

The aim of the work was to study the effect of milking fraction on electrical conductivity of milk (EC) to improve its use in dairy goat mastitis detection using automatic EC measurements during milking. The experiment was carried out on a group of 84 Murciano-Granadina goats (28 primiparous and 56 multiparous). Goats were in the fourth month of lactation. A linear mixed model was used to analyse the relationship between EC or somatic cell count (SCC) of gland milk and parity, mammary gland health status, analysed fraction (first 100 mL=F-1; machine milk=F-2; and stripping milk=F-3) and their first order interactions. Additionally, the mastitis detection characteristics (sensitivity, specificity, positive predictive value and negative predictive value) of SCC and EC were studied at different thresholds.All factors considered were significant for EC and SCC. EC decreased significantly as milking progressed (from F-1 to F-3) in both healthy and infected glands. EC was not significantly different between healthy and infected glands in F-1 and F-2 fractions, but EC of healthy glands (5.01 mS/cm) was significantly lower than in infected glands (5.03 mS/cm) at F-3.Mastitis detection characteristics of EC did not differ amongst studied fractions. The small significant difference of EC between healthy and infected glands obtained in F-3 fraction did not yield better sensitivity results compared to F-1 and F-2. The best EC mastitis detection characteristics were obtained at 5.20 mS/cm threshold (sensitivity of 70% and specificity of 50%). The best SCC mastitis detection characteristics were obtained at 300,000 cells/mL threshold and F-3 fraction (sensitivity of 85% and specificity of 65%).It was concluded that mastitis detection characteristics of EC were similar in the three milking fractions analysed, being slightly better for SCC in F-3 fraction. As shown in previous studies, there are no factors other than the mammary gland health status that affect milk EC and should be considered in the algorithms for mastitis detection to improve the results. (C) 2012 Elsevier B.V. All rights reserved.