Prediction of genetic gain from quadratic optimisation with constrained rates of inbreeding

There are selection methods available that allow the optimisation of genetic contributions of selection candidates for maximising the rate of genetic gain while restricting the rate of inbreeding. These methods imply selection on quadratic indices as the selection merit of a particular individual is a quadratic function of its estimated breeding value. This study provides deterministic predictions of genetic gain from selection on quadratic indices for a given set of resources (the number of candidates), heritability, and target rate of inbreeding. The rate of gain was obtained as a function of the accuracy of the Mendelian sampling term at the time of convergence of long-term contributions of selected candidates and the theoretical ideal rate of gain for a given rate of inbreeding after an exact allocation of long-term contributions to Mendelian sampling terms. The expected benefits from quadratic indices over traditional linear indices (i.e. truncation selection), both using BLUP breeding values, were quantified. The results clearly indicate higher gains from quadratic optimisation than from truncation selection. With constant rate of inbreeding and number of candidates, the benefits were generally largest for intermediate heritabilities but evident over the entire range. The advantage of quadratic indices was not highly sensitive to the rate of inbreeding for the constraints considered.     

Selection indices for non-linear profit functions

Summary Conventional selection index theory assumes that the total merit or profitability of animals is a linear function of measurable traits. However, in many cases merit may be a non-linear function of these traits. A linear selection index can still be used in this situation but the optimum index depends on the selection intensity to be used and on the number of generation over which the selection response is to be maximized. Nonlinear selection indices have been suggested but these result in a lower selection response than the best linear index. Linear selection indices suggested in the past are shown to correspond to the optimum linear index for either a very small selection response or, in the case of restricted indices, a very large selection response. The economic value of a trait may depend on management decisions taken by the farmer. In this situation the economic values should be calculated assuming that the management decisions taken maximize profit given the present genetic value of the animals.     

Computational optimal control for the time fractional convection-diffusion-reaction system

This paper proposes a numerical approximation method for computational optimal control of a time fractional convection-diffusion-reaction system. The proposed method involves discretizing the spatial domain by finite element method, approximating the admissible controls by control parameterization, and then obtaining an optimal parameter selection problem which can be solved by numerical optimization algorithms such as sequential quadratic programming. Specifically, an implicit finite difference method is employed to solve the time fractional system, and the sensitivity method for gradient computation in integer order optimal control problems is adjusted to the fractional order case. Simulation results demonstrate the validity and accuracy of the proposed numerical approximation method.     

Are economic selection indices always superior to a desired gains index?

Conclusions The comparison of different selection indices is justified only if the indices are constrated to achieve the same profit function, even when each index is not optimized with respect to that profit function.When a profit function is known and is non-linear, the desired gains index may be more efficient than the economic index. The optimum desired gains index should be determined by iterative techniques over several generations to compare the genetic progress with the economic index, because gains by the economic index are not linear and the changes observed in the initial generations of selection are not the same rates in future generations, although those changes are linear in the case of the desired gains index.     

Intersection–union tests for characterising recent changes in smoothed indicator time series

Using indicator time series for assessment and management requires methods for characterising recent time trends. We propose an approach where first the indicator time series is smoothed using a generalised additive model with optimal selection of the degree of smoothness. Second an intersection–union test is carried out using two test statistics which are the occurrence of the global maximum (or minimum) within the most recent years and the signs of the estimated annual first derivatives of the smoothed indicator times series during the same period, including years with missing data. The proposed test is applied to fish abundance indices for the North Sea, for which it is they are able to pick up changes happening during the last 3–5 years in contrast to linear regression and the Mann–Kendall test which find much fewer significant recent trends. An additional test for changes in trends using the second derivatives of the smoothed indicator time series provide early warnings for subsequent trends for certain species.     

Mendelian sampling terms as a selective advantage in optimum breeding schemes with restrictions on the rate of inbreeding

Quadratic indices are a general approach for the joint management of genetic gain and inbreeding in artificial selection programmes. They provide the optimal contributions that selection candidates should have to obtain the maximum gain when the rate of inbreeding is constrained to a predefined value. This study shows that, when using quadratic indices, the selective advantage is a function of the Mendelian sampling terms. That is, at all times, contributions of selected candidates are allocated according to the best available information about their Mendelian sampling terms (i.e. about their superiority over their parental average) and not on their breeding values. By contrast, under standard truncation selection, both estimated breeding values and Mendelian sampling terms play a major role in determining contributions. A measure of the effectiveness of using genetic variation to achieve genetic gain is presented and benchmark values of 0.92 for quadratic optimisation and 0.5 for truncation selection are found for a rate of inbreeding of 0.01 and a heritability of 0.25.     

A marginalized conditional linear model for longitudinal binary data when informative dropout occurs in continuous time

Within the pattern-mixture modeling framework for informative dropout, conditional linear models (CLMs) are a useful approach to deal with dropout that can occur at any point in continuous time (not just at observation times). However, in contrast with selection models, inferences about marginal covariate effects in CLMs are not readily available if nonidentity links are used in the mean structures. In this article, we propose a CLM for long series of longitudinal binary data with marginal covariate effects directly specified. The association between the binary responses and the dropout time is taken into account by modeling the conditional mean of the binary response as well as the dependence between the binary responses given the dropout time. Specifically, parameters in both the conditional mean and dependence models are assumed to be linear or quadratic functions of the dropout time; and the continuous dropout time distribution is left completely unspecified. Inference is fully Bayesian. We illustrate the proposed model using data from a longitudinal study of depression in HIV-infected women, where the strategy of sensitivity analysis based on the extrapolation method is also demonstrated.     

Constructing a confidence interval for the fraction who benefit from treatment,using randomized trial data

The fraction who benefit from treatment is the proportion of patients whose potential outcome under treatment is better than that under control. Inference on this parameter is challenging since it is only partially identifiable, even in our context of a randomized trial. We propose a new method for constructing a confidence interval for the fraction, when the outcome is ordinal or binary. Our confidence interval procedure is pointwise consistent. It does not require any assumptions about the joint distribution of the potential outcomes, although it has the flexibility to incorporate various user‐defined assumptions. Our method is based on a stochastic optimization technique involving a second‐order, asymptotic approximation that, to the best of our knowledge, has not been applied to biomedical studies. This approximation leads to statistics that are solutions to quadratic programs, which can be computed efficiently using optimization tools. In simulation, our method attains the nominal coverage probability or higher, and can have narrower average width than competitor methods. We apply it to a trial of a new intervention for stroke.     

Estimated quadratic inference function for correlated failure time data

An estimated quadratic inference function method is proposed for correlated failure time data with auxiliary covariates. The proposed method makes efficient use of the auxiliary information for the incomplete exposure covariates and preserves the property of the quadratic inference function method that requires the covariates to be completely observed. It can improve the estimation efficiency and easily deal with the situation when the cluster size is large. The proposed estimator which minimizes the estimated quadratic inference function is shown to be consistent and asymptotically normal. A chi-squared test based on the estimated quadratic inference function is proposed to test hypotheses about the regression parameters. The small-sample performance of the proposed method is investigated through extensive simulation studies. The proposed method is then applied to analyze the Study of Left Ventricular Dysfunction (SOLVD) data as an illustration.     

Adjusting for publication bias in meta-analysis via inverse probability weighting using clinical trial registries

Publication bias is a major concern in conducting systematic reviews and meta-analyses. Various sensitivity analysis or bias-correction methods have been developed based on selection models, and they have some advantages over the widely used trim-and-fill bias-correction method. However, likelihood methods based on selection models may have difficulty in obtaining precise estimates and reasonable confidence intervals, or require a rather complicated sensitivity analysis process. Herein, we develop a simple publication bias adjustment method by utilizing the information on conducted but still unpublished trials from clinical trial registries. We introduce an estimating equation for parameter estimation in the selection function by regarding the publication bias issue as a missing data problem under the missing not at random assumption. With the estimated selection function, we introduce the inverse probability weighting (IPW) method to estimate the overall mean across studies. Furthermore, the IPW versions of heterogeneity measures such as the between-study variance and the I² measure are proposed. We propose methods to construct confidence intervals based on asymptotic normal approximation as well as on parametric bootstrap. Through numerical experiments, we observed that the estimators successfully eliminated bias, and the confidence intervals had empirical coverage probabilities close to the nominal level. On the other hand, the confidence interval based on asymptotic normal approximation is much wider in some scenarios than the bootstrap confidence interval. Therefore, the latter is recommended for practical use.     

Nonparametric mixed effects models for unequally sampled noisy curves

We propose a method of analyzing collections of related curves in which the individual curves are modeled as spline functions with random coefficients. The method is applicable when the individual curves are sampled at variable and irregularly spaced points. This produces a low-rank, low-frequency approximation to the covariance structure, which can be estimated naturally by the EM algorithm. Smooth curves for individual trajectories are constructed as best linear unbiased predictor (BLUP) estimates, combining data from that individual and the entire collection. This framework leads naturally to methods for examining the effects of covariates on the shapes of the curves. We use model selection techniques--Akaike information criterion (AIC), Bayesian information criterion (BIC), and cross-validation--to select the number of breakpoints for the spline approximation. We believe that the methodology we propose provides a simple, flexible, and computationally efficient means of functional data analysis.     

Empirical comparison study of approximate methods for structure selection in binary graphical models

Looking for associations among multiple variables is a topical issue in statistics due to the increasing amount of data encountered in biology, medicine, and many other domains involving statistical applications. Graphical models have recently gained popularity for this purpose in the statistical literature. In the binary case, however, exact inference is generally very slow or even intractable because of the form of the so‐called log‐partition function. In this paper, we review various approximate methods for structure selection in binary graphical models that have recently been proposed in the literature and compare them through an extensive simulation study. We also propose a modification of one existing method, that is shown to achieve good performance and to be generally very fast. We conclude with an application in which we search for associations among causes of death recorded on French death certificates.     

A neural network approach to approximating MAP in belief networks

Bayesian belief networks (BBN) are a widely studied graphical model for representing uncertainty and probabilistic interdependence among variables. One of the factors that restricts the model's wide acceptance in practical applications is that the general inference with BBN is NP-hard. This is also true for the maximum a posteriori probability (MAP) problem, which is to find the most probable joint value assignment to all uninstantiated variables, given instantiation of some variables in a BBN. To circumvent the difficulty caused by MAP's computational complexity, we suggest in this paper a neural network approximation approach. With this approach, a BBN is treated as a neural network without any change or transformation of the network structure, and the node activation functions are derived based on an energy function defined over a given BBN. Three methods are developed. They are the hill-climbing style discrete method, the simulated annealing method, and the continuous method based on the mean field theory. All three methods are for BBN of general structures, with the restriction that nodes of BBN are binary variables. In addition, rules for applying these methods to noisy-or networks are also developed, which may lead to more efficient computation in some cases. These methods' convergence is analyzed, and their validity tested through a series of computer experiments with two BBN of moderate size and complexity. Although additional theoretical and empirical work is needed, the analysis and experiments suggest that this approach may lead to effective and accurate approximation for MAP problems.     

Method of decomposing spectra using the original spectrum and its derivatives

A method for the decomposition of optical spectra into bands was proposed, which is based on simultaneous approximation of the initial spectrum and its derivatives. The bands of the standard (gaussian) form were used in the decomposition procedure. A method for the selection of the optimal smoothing filter was described. The efficiency of the proposed method was demonstrated on model and experimental absorption spectra. It was shown that this method makes it possible to determine the number of bands and its parameters more exactly than the standard approach based on the analysis of one derivative of one power.     

Robust Gaussian Graphical Modeling Via l1 Penalization

Summary Gaussian graphical models have been widely used as an effective method for studying the conditional independency structure among genes and for constructing genetic networks. However, gene expression data typically have heavier tails or more outlying observations than the standard Gaussian distribution. Such outliers in gene expression data can lead to wrong inference on the dependency structure among the genes. We propose a l₁ penalized estimation procedure for the sparse Gaussian graphical models that is robustified against possible outliers. The likelihood function is weighted according to how the observation is deviated, where the deviation of the observation is measured based on its own likelihood. An efficient computational algorithm based on the coordinate gradient descent method is developed to obtain the minimizer of the negative penalized robustified‐likelihood, where nonzero elements of the concentration matrix represents the graphical links among the genes. After the graphical structure is obtained, we re‐estimate the positive definite concentration matrix using an iterative proportional fitting algorithm. Through simulations, we demonstrate that the proposed robust method performs much better than the graphical Lasso for the Gaussian graphical models in terms of both graph structure selection and estimation when outliers are present. We apply the robust estimation procedure to an analysis of yeast gene expression data and show that the resulting graph has better biological interpretation than that obtained from the graphical Lasso.     

Adjustment of Prediction Intervals in Nonlinear Regression

By treating the nonlinear model as if it were linear in the parameterization θ in the neighbourhood of the least squares estimate θC, two-sided nominally-q-prediction intervals can be constructed by applying the usual linear model theory. The quadratic approximation of the expected coverage of the prediction intervals is derived for a p-parameter nonlinear model. An adjustment of the nominally-q-prediction intervals is proposed. It is shown that, to the extent that quadratic approximation is adequate, the actual expected coverage of the adjusted prediction intervals is q.     

Chemotherapy for tumors: an analysis of the dynamics and a study of quadratic and linear optimal controls

We investigate a mathematical model of tumor-immune interactions with chemotherapy, and strategies for optimally administering treatment. In this paper we analyze the dynamics of this model, characterize the optimal controls related to drug therapy, and discuss numerical results of the optimal strategies. The form of the model allows us to test and compare various optimal control strategies, including a quadratic control, a linear control, and a state-constraint. We establish the existence of the optimal control, and solve for the control in both the quadratic and linear case. In the linear control case, we show that we cannot rule out the possibility of a singular control. An interesting aspect of this paper is that we provide a graphical representation of regions on which the singular control is optimal.     

Analysis der Basalstoffwechselregelung in Abhängigkeit von Umweltfaktoren und Körpergewicht

In large-scale experiments carried out over a period of about 15 years, white rats were subjected to chronic treatment with antithyroid drugs, substances counteractingLangerhans' islets (mostly alloxan), hormone fragments, and native hormones respectively. The O₂-consumption (mostly in an artificial microclimate) was measured in each test. The test batches were either gradually diminished by killing the animals one after another, to yield histological series in chronological order, or a whole batch was pooled together at the end of a test. In all experiments a relationship could be established between O₂-consumption and the histology (histochemical tests included) of liver and endocrine glands. In a recent experiment, moreover, serum enzymes and the level of blood sugar were determined; a chromatographic analysis of thyroid hormone fractions and blood cells-counts were made. Combining these methods led to numerous variables suitable for numerical treatment. Besides a treatment of the data by elementary methods (variance analysis, t-test, U-test, linear 2-dimensional regression analysis), multidimensional space models, employing traditional and novel calculi were used. The regression equations in 23-dimensional statements were linear, quadratic and cubic. In a concordance analysis 25 variables were carried along. To develop a satisfactory theory, the coefficients of the differential equations underlying the material were determined by means of an approximation method suited for random samples subject to variance. For solution, integrals, polynomials of e-functions, using mostly the regression calculus in two steps were sought. The oscillations caused by environmental factors were extracted as the differences between the function values of the optimal particular and real integral and the empirical ordinates. This isolated swinging remainder was analysed followingFourier. The spectra or the oscillation portions themselves were brought into functional relation with one another by means of functional operators always controlled by hetero- and auto-correlation. The model of the control of metabolism by the diencephalic-pituitary-thyroid-system could thus be reduced to an n-dimensional linear space model, which, however, is not quite satisfactory.     

A class of multiplicity adjusted tests for spatial clustering based on case-control point data

A class of tests with quadratic forms for detecting spatial clustering of health events based on case-control point data is proposed. It includes Cuzick and Edwards's test statistic (1990, Journal of the Royal Statistical Society, Series B 52, 73-104). Although they used the property of asymptotic normality of the test statistic, we show that such an approximation is generally poor for moderately large sample sizes. Instead, we suggest a central chi-square distribution as a better approximation to the asymptotic distribution of the test statistic. Furthermore, not only to estimate the optimal value of the unknown parameter on the scale of cluster but also to adjust for multiple testing due to repeating the procedure by changing the parameter value, we propose the minimum of the profile p-value of the test statistic for the parameter as an integrated test statistic. We also provide a statistic to estimate the areas or cases which make large contributions to significant clustering. The proposed methods are illustrated with a data set concerning the locations of cases of childhood leukemia and lymphoma and another on early medieval grave site locations consisting of affected and nonaffected grave sites.     

Methods of selecting informative variables

We propose a new method for selection of the most informative variables from the set of variables which can be measured directly. The information is measured by metrics similar to those used in experimental design theory, such as determinant of the dispersion matrix of prediction or various functions of its eigenvalues. The basic model admits both population variability and observational errors, which allows us to introduce algorithms based on ideas of optimal experimental design. Moreover, we can take into account cost of measuring various variables which makes the approach more practical. It is shown that the selection of optimal subsets of variables is invariant to scale transformations unlike other methods of dimension reduction, such as principal components analysis or methods based on direct selection of variables, for instance principal variables and battery reduction. The performance of different approaches is compared using the clinical data.