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.

Comparação de ajustes do modelo de Gompertz a dados de crescimento

Pós-graduação em Biometria - IBB

Liquid nitrogen for the treatment of actinic keratosis: A longitudinal assessment

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Efeitos da adição do IGF-1 ou IGF-LongR3 sobre aspectos celulares e moleculares de complexos cumulus-oócito durante a maturação oocitária in vitro em bovinos

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

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.

Avaliação dos desempenhos produtivo e reprodutivo de um rebanho bubalino no Estado do Ceará

Foram avaliados os desempenhos produtivo e reprodutivo de um rebanho bubalino, localizado no município de Paracuru, litoral oeste do Estado do Ceará. Os animais foram mantidos em sistema intensivo e submetidos a duas ordenhas diárias. Foram analisadas informações de 87 fêmeas da raça Murrah, entre os anos de 1984 e 1998. As análises estatísticas foram efetuadas por meio do procedimento GLM (SAS, 1990), usando um modelo linear misto. As médias observadas para produção total de leite, duração da lactação, produção de leite no pico de lactação, idade ao primeiro parto e intervalo de partos foram: 2130,80+535,60kg, 301,41+49,30dias, 9,78+1,95kg, 1132,69+166,99dias e 430,79+100,44dias, respectivamente. Alta porcentagem de parições das ocorreu no período chuvoso (79,3%) e apenas 20,7%, no período seco. A produção total de leite e a produção no pico de lactação foram influenciadas pelos efeitos do ano e ordem de parto. A duração da lactação foi influenciada pelos efeitos do ano e mês de parto. A idade ao primeiro parto foi influenciada pelo efeito do ano de nascimento da búfala. O intervalo de partos foi influenciado pelo mês do parto anterior, além dos efeitos do mês dentro do ano de parto e mês de nascimento da búfala.

Variação genética em progênies de Myracrodruon urundeuva F.F. & M.F. Allemão em três sistemas de cultivo

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

Narrativas orais de crianças com desenvolvimento típico de linguagem

TEMA: desenvolvimento da narrativa oral. OBJETIVO: verificar o tempo de narrativa e de pausa, o número de palavras e de intervenções do interlocutor em narrativas orais de crianças com desenvolvimento típico. MÉTODO: participaram do estudo 31 crianças divididas em quatro grupos etários: GI (3:1 a 4:0 anos), GII (4:1 a 5:0 anos), GIII (5:1 a 6:0 anos) e GIV (6:1 a 7:0 anos). Amostras de narrativa espontânea e narrativa com livro sem palavras foram coletadas em vídeo, transcritas e analisadas estatisticamente por meio de teste exato de Fisher (não-paramétrico) e modelo de regressão linear com efeitos mistos. RESULTADOS: os valores de tempo de pausa, tempo de narrativa, e o número de palavras no contexto de livro foram significativamente maiores em relação à narrativa espontânea (p-valor < 0,01). Quanto ao número de intervenções, houve correlação (p-valor = 0,03) entre idade e intervenção no contexto de livro com diminuição da intervenção na media que aumentou a idade. CONCLUSÃO: as crianças apresentaram uma narrativa mais extensa no contexto de relato com livro sem palavras em relação ao contexto de narrativa espontânea, porém, sem diferenças significativas entre as idades. O estudo permitiu ainda concluir que a participação do interlocutor faz-se menos necessária conforme aumenta a idade do narrador.

Use of survival analysis as a tool for the genetic improvement of age at first conception in Nellore cattle

Restricted breeding seasons in beef cattle lead to censoring of reproductive data. In this paper, age at first conception (AFC) of Nellore females exposed to the sires for the first time between 11 and 16 months of age, was studied aiming to verify the possibility of genetically advance sexual precocity using a survival model. The final data set contained 6699 records of AFC in days. Records of females that did not calve in the next year following exposure to the sire were considered censored (77.5% of total). The model used was a Weibull mixed survival model including effects of contemporary groups, period (fixed) and animal (random). The effect of the contemporary groups on AFC was important (p < 0.01). Heritabilities were 0.51 and 0.76 in logarithmic and original scales respectively. Results indicate that it is possible to genetically advance sexual precocity, using the outcome of survival analysis of AFC as selection criterion. They also suggest that improvements of the environment could advance sexual precocity too, thus an adequate pregnancy rate for farmers could quickly be achieved.

The effect of adhesive and flowable composite on postoperative sensitivity: 2-Week results

Objective: To measure 2-week postoperative sensitivity in Class II composite restorations placed with a self-etching adhesive (Clearfil SE Bond) or a total-etch adhesive (Prime&Bond NT) with or without a flowable composite as cervical increment. Method and materials: Upon approval by the University of Guarulhos Committee on Human Subjects, 100 restorations were inserted in 46 patients who required Class II restorations in their molars and premolars. Enamel and dentin walls were conditioned with a self-etching primer (for Clearfil SE Bond) or etched with 34% phosphoric acid (for Prime&Bond NT). A 1- to 2-mm-thick increment of a flowable composite (Filtek Flow) was used in the proximal box in 50% of the restorations of each adhesive. Preparations were restored with a packable composite (Surefil). The restorations were evaluated preoperatively and 2 weeks postoperatively for sensitivity to cold, air, and masticatory forces using a visual analog scale. Marginal integrity of the accessible margins was also evaluated. Statistical analysis used a mixed linear model with subject as a random effect. Results: Ninety-eight teeth from 44 subjects were observed at 2 weeks. The type of adhesive and use of flowable composite had no significant effects or interaction for any of the four outcomes of interest, ie, change from baseline to 2 weeks in sensitivity and response time for the cold or air stimulus. For the air stimulus, the overall average change from baseline was not significant for either sensitivity or response time. For the cold stimulus, the overall average change from baseline was significant for both sensitivity and response time. No case of sensitivity to masticatory forces was observed. Conclusion: No differences in postoperative sensitivity were observed between a self-etch adhesive and a total-etch adhesive at 2 weeks. The use of flowable composite did not decrease postoperative sensitivity.

Environmental factors influencing foraging activity in the social wasp Polybia paulista (Hymenoptera: Vespidae: Epiponini)

Foraging behavior in social wasps is important in the development of the colony and reflects an important ecological interaction between the colony and the environment. Although the social traits of the colony play a role in the foraging activities, the conditions that establish the space and time limits are mainly physical. Here, we evaluate colonies of Polybia paulista throughout one year in order to verify the foraging activities and the items collected, as well as the importance of temperature, relative humidity, and solar radiation on motivating foraging. Collection of liquids was always higher than that of solids; preys were collected all year long, and nests showed two annual episodic expansions. The linear mixed effects (LME) model used to analyze which weather factors influence the foraging showed temperature as the most influencing factor on the collection of materials. © 2011 Naila Cristina de Souza Canevazzi and Fernando Barbosa Noll.

Strategies to model scheduling decisions to plan the soft drink production process

In this paper we present a mixed integer model that integrates lot sizing and lot scheduling decisions for the production planning of a soft drink company. The main contribution of the paper is to present a model that differ from others in the literature for the constraints related to the scheduling decisions. The proposed strategy is compared to other strategies presented in the literature.

Delineamento ótimos para modelos não-lineares: efeitos fixos e efeitos aleatórios

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Quantitative genetic study of age at subsequent rebreeding in Nellore cattle by using survival analysis

The continuous trait age at subsequent rebreeding (ASR) was evaluated using survival analysis in Nellore breed cows that conceived for the first time at approximately 14 months of age. This methodology was chosen because the restricted breeding season produces censored data. The dataset contained 2885 records of ASR (in days). Records of females that did not produce calves in the following year after being exposed to a sire were considered censored (48.3% of the total). The statistical model used was a Weibull mixed survival model, which included fixed effects of contemporary groups (CG) and period and a random effect of individual animal. The effect of contemporary groups on ASR was significant (P < 0.01). Heritabilities obtained for ASR were 0.03 and 0.04 in logarithmic and original scales, respectively. These results indicate that the genetic selection response for subsequent reproduction of 2-year-old Nellore breed females is not expected to be effective based on survival analysis. Furthermore, these results suggest that environmental improvement is fundamental to this important trait. It should be highlighted that an increase in the average date of birth can produce an adverse effect in the future, since this cannot be compensated by genetic improvement.

Ajustes de modelos não lineares de efeitos fixos, com ponderação e misto

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)