The Treatment of Missing Values in Logistic Regression

The efficiencies of the estimators in the linear logistic regression model are examined using simulations under six missing value treatments. These treatments use either the maximum likelihood or the discriminant function approach in the estimation of the regression coefficients. Missing values are assumed to occur at random. The cases of multivariate normal and dichotomous independent variables are both considered. We found that in general, there is no uniformly best method. However, mean substitution and discriminant function estimation using existing pairs of values for correlations turn out to be favourable for the cases considered.     

The Monofactorial Model for Recurrence Risks with Generalized Penetrance

Recurrence risks are derived explicitly in terms of gene frequencies and penetrance coefficients for the general case in which all genotypes have incomplete penetrance. Maximum likelihood estimation of recurrence risks is achieved through the use of the semi-symmetric intraclass contingency table. The resulting formulas and estimation procedure can be useful for the analysis of population and family data, and in genetic counselling.     

Describing Ordinal Odds Ratios for Stratified r × c Tables

For an r × ctable with ordinal responses, odds ratios are commonly used to describe the relationship between the row and column variables. This article shows two types of ordinal odds ratios where local‐global odds ratios are used to compare several groups on a c‐category ordinal response and a global odds ratio is used to measure the global association between a pair of ordinal responses. When there is a stratification factor, we consider Mantel‐Haenszel (MH) type estimators of these odds ratios to summarize the association from several strata. Like the ordinary MH estimator of the common odds ratio for several 2 × 2 contingency tables, the estimators are used when the association is not expected to vary drastically among the strata. Also, the estimators are consistent under the ordinary asymptotic framework in which the number of strata is fixed and also under sparse asymptotics in which the number of strata grows with the sample size. Compared to the maximum likelihood estimators, simulations find that the MH type estimators perform better especially when each stratum has few observations. This article provides variances and covariances formulae for the local‐global odds ratios estimators and applies the bootstrap method to obtain a standard error for the global odds ratio estimator. At the end, we discuss possible ways of testing the homogeneity assumption.     

Introducing the Multivariate Dale Model in Population‐Based Genetic Association Studies

Until recently, the most common parametric approaches to study the combined effects of several genetic polymorphisms located within a gene or in a small genomic region are, at the genotype level, logistic regressions and at the haplotype level, haplotype analyses. An alternative modeling approach, based on the case/control principle, is to regard exposures (e.g., genetic data such as derived from Single Nucleotide Polymorphisms – SNPs) as random and disease status as fixed and to use a marginal multivariate model that accounts for inter‐relationships between exposures. One such model is the multivariate Dale model. This model is based on multiple logistic regressions. That is why the model, applied in a case/control setting, leads to straightforward interpretations that are similar to those drawn in a classical logistic modeling framework. (© 2004 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Gramicidin A-phospholipid model systems

Gramicidin A forms ion-conducting channels which can traverse the hydrocarbon core of lipid bilayer membranes. The structures formed by gramicidin A are among the best characterized of all membrane-bound polypeptides or proteins. In this review a brief summary is given of the occurrence, conformation, and synthesis of gramicidin A, and of its use as a model for ion transport and the interaction of proteins and lipids in biological membranes.     

A Markov approach for describing post-fire succession of vegetation

Vegetation dynamics in the coastal area of the Seto Inland Sea region in Japan, where wild fires occur frequently, were described using a stationary Markov model. In this region, vegetation types ofMiscanthus-Pleioblastus grassland,Lespedeza-Mallotus scrub,Pinus-Rhododendron forest andCrassocephalum-Erechtites community have been identified, and these show cyclic succession under the influence of fires. The model uses parameters determining fire frequency and rate of successional change to analyze the effect of variation in these parameters on the areal ratio of each vegetation type at equilibrium and on the time taken for one vegetation type to succeed another (elapsed successional time). The effect of fire frequency differs between hypothetical habitats with high and low productivity. A policy for vegetation management in areas of high and low productivity is proposed. The advantages and limitations of applying Markov models to studies of vegetation succession are also discussed.     

Diffusion of inorganic carbon across an unstirred layer: a simplified quantitative approach

Abstract The quantitative approach used here is based on a model comprising a well-stirred medium, an unstirred layer, and a CO₂ absorbing leaf. The unstirred layer is divided up by planes into a number of sub-layers. Within each plane the concentration of each solute is everywhere the same as is the electric potential. These variables constitute the basic data. Thus the planes were characterized by their pH value. An equation is derived which enables the calculation of the basic data of a plane from the known data of another plane. In this way it is possible to calculate the basic data for all planes. From these data the rate of assimilation, the thickness of the unstirred layer and its sub-layers, the fluxes across the sub-layers and the conversions among the carbon components can be estimated. The CO₂ flux decreases, and the HCO^?₃ flux increases towards the leaf. There are negative fluxes of OH^& and CO^2–₃. H⁺ fluxes are of minor importance and can be ignored if the pH of the medium is higher than 8.0, provided no non-inorganic C buffers with appropriate pK_a are present. The significance of the carbon diffusion facilitating effect of an inorganic carbon system is expressed in various ways. The values obtained represent maxima, as the assumption is made that the equilibrium reactions are very fast. It is argued that even better effects are possible if the back-diffusion of CO^2–₃ could be prevented by lowering the pH of the unstirred layer.