A maximum likelihood framework for protein design

Affiliation: Claudia Kleinman, Nicolas Rodrigue & Hervé Philippe : Département de biochimie, Faculté de médecine, Université de Montréal

Revisiting optimization algorithms for maximum likelihood estimation

Parmi les méthodes d’estimation de paramètres de loi de probabilité en statistique, le maximum de vraisemblance est une des techniques les plus populaires, comme, sous des conditions l´egères, les estimateurs ainsi produits sont consistants et asymptotiquement efficaces. Les problèmes de maximum de vraisemblance peuvent être traités comme des problèmes de programmation non linéaires, éventuellement non convexe, pour lesquels deux grandes classes de méthodes de résolution sont les techniques de région de confiance et les méthodes de recherche linéaire. En outre, il est possible d’exploiter la structure de ces problèmes pour tenter d’accélerer la convergence de ces méthodes, sous certaines hypothèses. Dans ce travail, nous revisitons certaines approches classiques ou récemment d´eveloppées en optimisation non linéaire, dans le contexte particulier de l’estimation de maximum de vraisemblance. Nous développons également de nouveaux algorithmes pour résoudre ce problème, reconsidérant différentes techniques d’approximation de hessiens, et proposons de nouvelles méthodes de calcul de pas, en particulier dans le cadre des algorithmes de recherche linéaire. Il s’agit notamment d’algorithmes nous permettant de changer d’approximation de hessien et d’adapter la longueur du pas dans une direction de recherche fixée. Finalement, nous évaluons l’efficacité numérique des méthodes proposées dans le cadre de l’estimation de modèles de choix discrets, en particulier les modèles logit mélangés.

Improved maximum likelihood estimators in a heteroskedastic errors-in-variables model

This paper develops a bias correction scheme for a multivariate heteroskedastic errors-in-variables model. The applicability of this model is justified in areas such as astrophysics, epidemiology and analytical chemistry, where the variables are subject to measurement errors and the variances vary with the observations. We conduct Monte Carlo simulations to investigate the performance of the corrected estimators. The numerical results show that the bias correction scheme yields nearly unbiased estimates. We also give an application to a real data set.

A matrix formula for the skewness of maximum likelihood estimators

We give a general matrix formula for computing the second-order skewness of maximum likelihood estimators. The formula was firstly presented in a tensorial version by Bowman and Shenton (1998). Our matrix formulation has numerical advantages, since it requires only simple operations on matrices and vectors. We apply the second-order skewness formula to a normal model with a generalized parametrization and to an ARMA model. (c) 2010 Elsevier B.V. All rights reserved.

Improved maximum-likelihood estimation in a regression model with general parametrization

We analyse the finite-sample behaviour of two second-order bias-corrected alternatives to the maximum-likelihood estimator of the parameters in a multivariate normal regression model with general parametrization proposed by Patriota and Lemonte [A. G. Patriota and A. J. Lemonte, Bias correction in a multivariate regression model with genereal parameterization, Stat. Prob. Lett. 79 (2009), pp. 1655-1662]. The two finite-sample corrections we consider are the conventional second-order bias-corrected estimator and the bootstrap bias correction. We present the numerical results comparing the performance of these estimators. Our results reveal that analytical bias correction outperforms numerical bias corrections obtained from bootstrapping schemes.

Análise Genética de Características Reprodutivas Na Raça Brahman

Aiming to establish selection criteria for the Brahman cattle in Brazil, the objective of this paper was to estimate genetic and environmental parameters for birth weight (BW), gestation length (GL) and scrotal circumference at 365 (SC365), 455 (SC455) and 550 (SC 550) days of age, and relative efficiency of indirect versus direct selection for these traits. Data were obtained from 15 851 animals belonging to 16 herds in six states, enrolled in a genetic breeding program. (Co)variance and heritability coefficients and genetic and environmental correlations between traits were estimated by restricted maximum likelihood (REML) under a multi trait animal model. The model included as fixed effects the contemporary group and the age of cow at calving as a covariate (linear and quadratic), and as random the direct genetic, the permanent environmental and residual effects. For BW and GL it was also considered the maternal genetic effect as random. Estimates of direct heritabilities were 0.36, 0.29, 0.55, 0.43 and 0.40 for BW, GL, SC365, SC455 and SC550, respectively. Genetic correlations between BW and GL, GL and SC365, GL and SC455, GL and SC550, SC365 and SC455, SC365 and SC550 and SC455 and SC550, were respectively 0.06, 0.13, 0.20, 0.13, 0.96, 0.98 and 0.99. The heritabilities estimates for all traits indicated these show enough additive genetic variability to respond favorably to selection. For sexual precocity, the best option would be SC365, due to the high value of heritability and to indirect selection at this age being more efficient than direct selection for PE455.

String-averaging expectation-maximization for maximum likelihood estimation in emission tomography

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

Application of Reactions Norms in Study of Genotype Environmental Interaction for Milk Yield of Buffaloes

The aim of this study was to estimate genetic parameters for milk yield (MY) in buffaloes using reaction norms. Model included the additive direct effect as random and contemporary group (herd and year of birth) were included as fixed effects and cow age classes (linear) as covariables. The animal additive direct random effect was modeled through linear Legendre polynomials on environment gradient (EG) standardized means. Mean trends were taken into account by a linear regression on Legendre polynomials of environmental group means. Residual variance was modeled trough 6 heterogeneity classes (EG). These classes of residual variance was formed : EG1: mean = 866,93 kg (621,68 kg-1011,76 kg); EG2: mean = 1193,00 kg (1011,76 kg-1251,49 kg); EG3: mean = 1309,37 kg (1251,49 kg -1393,20 kg); EG4: mean = 1497,59 kg (1393,20 kg-1593,53 kg); EG5: mean = 1664,78 kg (1593,53 kg -1727,32kg) e EG6: mean = 1973,85 kg (1727,32 kg -2422,19 kg).(Co) variance functions were estimated by restricted maximum likelihood (REML) using the GIBBS3F90 package. The heritability estimates for MY raised as the environmental gradient increased, varying from 0.20 to 0.40. However, in intermediate to favorable environments, the heritability estimates obtained with Considerable genotype-environment interaction was found for MY using reaction norms. For genetic evaluation of MY is necessary to consider heterogeneity of variances to model the residual variance.