Copulas in QTL mapping

The standard variance components method for mapping quantitative trait loci is derived on the assumption of normality. Unsurprisingly, statistical tests based on this method do not perform so well if this assumption is not satisfied. We use the statistical concept of copulas to relax the assumption of normality and derive a test that can perform well under any distribution of the continuous trait. In particular, we discuss bivariate normal copulas in the context of sib-pair studies. Our approach is illustrated by a linkage analysis of lipoprotein(a) levels, whose distribution is highly skewed. We demonstrate that the asymptotic critical levels of the test can still be calculated using the interval mapping approach. The new method can be extended to more general pedigrees and multivariate phenotypes in a similar way as the original variance components method.

Nutrient partitioning during treatment of tuberculosis: gain in body fat mass but not in protein mass

Background: Tuberculosis is an important cause of wasting. The functional consequences of wasting and recovery may depend on the distribution of lost and gained nutrient stores between protein and fat masses. Objective: The goal was to study nutrient partitioning, ie, the proportion of weight change attributable to changes in fat mass (FM) versus protein mass (PM), during anti mycobacterial treatment. Design: Body-composition measures were made of 21 men and 9 women with pulmonary tuberculosis at baseline and after 1 and 6 mo of treatment. All subjects underwent dual-energy X-ray absorptiometry and deuterium bromide dilution tests, and a four-compartment model of FM, total body water (TBW), bone minerals (BM), and PM was derived. The ratio of PM to FM at any time was expressed as the energy content (p-ratio). Changes in the p-ratio were related to disease severity as measured by radiologic criteria. Results: Patients gained 10% in body weight (P < 0.001) from baseline to month 6. This was mainly due to a 44% gain in FM (P < 0.001); PM, BM, and TBW did not change significantly. Results were similar in men and women. The p-ratio decreased from baseline to month 1 and then fell further by month 6. Radiologic disease severity was not correlated with changes in the p-ratio. Conclusions: Microbiological cure of tuberculosis does not restore PM within 6 mo, despite a strong anabolic response. Change in the p-ratio is a suitable parameter for use in studying the effect of disease on body composition because it allows transformation of such effects into a normal distribution across a wide range of baseline proportion between fat and protein mass.

Virus-specific CD8+ T lymphocytes within the normal human liver

The frequency and phenotype of human antiviral memory CD8(+) T cells in blood are well studied, yet little is known about their distribution within tissues. Analysis of antiviral CD8(+) T cell populations derived from a unique set of normal liver and blood samples identified a consistent population of virus-specific cells within the liver. In comparison to the circulating T cells, the liver-derived T cells were present at frequencies which were variably enriched compared to that in the blood, and showed significant differences with regard to the expression of CD45RA, CD45RO, CD95, CCR7, CD27 and CD28. The differences in these cell surface markers are consistent with a mature 'effector memory' phenotype of antigen-specific CD8(+) T cells within the liver. An enrichment of an activated subset of NKT cells (Valpha24/Vbeta11) was also observed, a finding which may be relevant to the regulation of the antiviral population:.

Hepatic pharmacokinetics of taurocholate in the normal and cholestatic rat liver

1 The disposition kinetics of [H-3] taurocholate ([H-3]TC) in perfused normal and cholestatic rat livers were studied using the multiple indicator dilution technique and several physiologically based pharmacokinetic models. 2 The serum biochemistry levels, the outflow profiles and biliary recovery of [H-3] TC were measured in three experimental groups: (i) control; (ii) 17α-ethynylestradiol (EE)-treated (low dose); and (iii) EE-treated (high dose) rats. EE treatment caused cholestasis in a dose-dependent manner. 3 A hepatobiliary TC transport model, which recognizes capillary mixing, active cellular uptake, and active efflux into bile and plasma described the disposition of [H-3]TC in the normal and cholestatic livers better than the other pharmacokinetic models. 4 An estimated five- and 18-fold decrease in biliary elimination rate constant, 1.7- and 2.7-fold increase in hepatocyte to plasma efflux rate constant, and 1.8- and 2.8-fold decrease in [H-3]TC biliary recovery ratio was found in moderate and severe cholestasis, respectively, relative to normal. 5 There were good correlations between the predicted and observed pharmacokinetic parameters of [H-3]TC based on liver pathophysiology (e.g. serum bilirubin level and biliary excretion of [H-3]TC). In conclusion, these results show that altered hepatic TC pharmacokinetics in cholestatic rat livers can be correlated with the relevant changes in liver pathophysiology in cholestasis.

Adsorption in slit pores and pore-size distribution: A molecular layer structure theory

A new approach is developed to analyze the thermodynamic properties of a sub-critical fluid adsorbed in a slit pore of activated carbon. The approach is based on a representation that an adsorbed fluid forms an ordered structure close to a smoothed solid surface. This ordered structure is modelled as a collection of parallel molecular layers. Such a structure allows us to express the Helmholtz free energy of a molecular layer as the sum of the intrinsic Helmholtz free energy specific to that layer and the potential energy of interaction of that layer with all other layers and the solid surface. The intrinsic Helmholtz free energy of a molecular layer is a function (at given temperature) of its two-dimensional density and it can be readily obtained from bulk-phase properties, while the interlayer potential energy interaction is determined by using the 10-4 Lennard-Jones potential. The positions of all layers close to the graphite surface or in a slit pore are considered to correspond to the minimum of the potential energy of the system. This model has led to accurate predictions of nitrogen and argon adsorption on carbon black at their normal boiling points. In the case of adsorption in slit pores, local isotherms are determined from the minimization of the grand potential. The model provides a reasonable description of the 0-1 monolayer transition, phase transition and packing effect. The adsorption of nitrogen at 77.35 K and argon at 87.29 K on activated carbons is analyzed to illustrate the potential of this theory, and the derived pore-size distribution is compared favourably with that obtained by the Density Functional Theory (DFT). The model is less time-consuming than methods such as the DFT and Monte-Carlo simulation, and most importantly it can be readily extended to the adsorption of mixtures and capillary condensation phenomena.

A simple implementation of a normal mixture approach to differential gene expression in multiclass microarrays

Motivation: An important problem in microarray experiments is the detection of genes that are differentially expressed in a given number of classes. We provide a straightforward and easily implemented method for estimating the posterior probability that an individual gene is null. The problem can be expressed in a two-component mixture framework, using an empirical Bayes approach. Current methods of implementing this approach either have some limitations due to the minimal assumptions made or with more specific assumptions are computationally intensive. Results: By converting to a z-score the value of the test statistic used to test the significance of each gene, we propose a simple two-component normal mixture that models adequately the distribution of this score. The usefulness of our approach is demonstrated on three real datasets.