Multifactor dimensionality reduction (MDR) was developed as a method for detecting statistical patterns of epistasis. The overall goal of MDR is to change the representation space of the data to make interactions easier to detect. It is well known that machine learning methods may not provide robust models when the class variable (e.g. case-control status) is imbalanced and accuracy is used as the fitness measure. This is because most methods learn patterns that are relevant for the larger of the two classes. The goal of this study was to evaluate three different strategies for improving the power of MDR to detect epistasis in imbalanced datasets. The methods evaluated were: (1) over-sampling that resamples with replacement the smaller class until the data are balanced, (2) under-sampling that randomly removes subjects from the larger class until the data are balanced, and (3) balanced accuracy [(sensitivity+specificity)/2] as the fitness function with and without an adjusted threshold. These three methods were compared using simulated data with two-locus epistatic interactions of varying heritability (0.01, 0.025, 0.05, 0.1, 0.2, 0.3, 0.4) and minor allele frequency (0.2, 0.4) that were embedded in 100 replicate datasets of varying sample sizes (400, 800, 1600). Each dataset was generated with different ratios of cases to controls (1 : 1, 1 : 2, 1 : 4). We found that the balanced accuracy function with an adjusted threshold significantly outperformed both over-sampling and under-sampling and fully recovered the power. These results suggest that balanced accuracy should be used instead of accuracy for the MDR analysis of epistasis in imbalanced datasets.

Human natural killer (NK) T cells are unique T lymphocytes that express an invariant T cell receptor (TCR) Vα24-Vβ11 and have been implicated to play a role in various diseases. A subset of NKT cells express CD4 and hence are potential targets for human immunodeficiency virus (HIV)-1 infection. We demonstrate that both resting and activated human Vα24+ T cells express high levels of the HIV-1 coreceptors CCR5 and Bonzo (CXCR6), but low levels of CCR7, as compared with conventional T cells. Remarkably NKT cells activated with α-galactosylceramide (α-GalCer)-pulsed dendritic cells were profoundly more susceptible to infection with R5-tropic, but not X4-tropic, strains of HIV-1, compared with conventional CD4+ T cells. Furthermore, resting CD4+ NKT cells were also more susceptible to infection. After initial infection, HIV-1 rapidly replicated and depleted the CD4+ subset of NKT cells. In addition, peripheral blood NKT cells were markedly and selectively depleted in HIV-1 infected individuals. Although the mechanisms of this decline are not clear, low numbers or absence of NKT cells may affect the course of HIV-1 infection. Taken together, our findings indicate that CD4+ NKT cells are directly targeted by HIV-1 and may have a potential role during viral transmission and spread in vivo.

These results provide further evidence for a role of RAAS activation in the pathophysiology of AF and point to a potential role for stratification of therapeutic approaches by ACE genotype.

This nested case-control study examined relationships between MDR1, CYP2B6, and CYP3A4 variants and hepatotoxicity during antiretroviral therapy with either efavirenz- or nevirapine-containing regimens. Decreased risk of hepatotoxicity was associated with MDR1 3435C-->T (odds ratio, 0.254; P=.021). An interaction between MDR1 and hepatitis B surface antigen status predicted risk with 82% accuracy (P<.001).

In the quest for discovering disease susceptibility genes, the reality of gene-gene and gene-environment interactions creates difficult challenges for many current statistical approaches. In an attempt to overcome limitations with current disease gene detection methods, the multifactor dimensionality reduction (MDR) approach was previously developed. In brief, MDR is a method that reduces the dimensionality of multilocus information to identify polymorphisms associated with an increased risk of disease. This approach takes multilocus genotypes and develops a model for defining disease risk by pooling high-risk genotype combinations into one group and low-risk combinations into another. Cross-validation and permutation testing are used to identify optimal models. While this approach was initially developed for studies of complex disease, it is also directly applicable to pharmacogenomic studies where the outcome variable is drug treatment response/nonresponse or toxicity/no toxicity. MDR is a nonparametric and model-free approach that has been shown to have reasonable power to detect epistasis in both theoretical and empirical studies. This computational technology is described in detail in this review, and its application in pharmacogenomic studies is demonstrated.

Genome-wide association studies have become a reality in the study of the genetics of complex disease. This technology provides a wealth of genomic information on patient samples, from which we hope to learn novel biology and detect important genetic and environmental factors for disease processes. Because strategies for analyzing these data have not kept pace with the laboratory methods that generate the data it is unlikely that these advances will immediately lead to an improved understanding of the genetic contribution to common human disease and drug response. Currently, no single analytical method will allow us to extract all information from a whole-genome association study. Thus, many novel methods are being proposed and developed. It will be vital for the success of these new methods, to have the ability to simulate datasets consisting of polymorphisms throughout the genome with realistic linkage disequilibrium patterns. Within these datasets, we can embed genetic models of disease whereby we can evaluate the ability of novel methods to detect these simulated effects. This paper describes a new software package, genomeSIM, for the simulation of large-scale genomic data in population based case-control samples. It allows for single SNP, as well as gene-gene interaction models to be associated with disease risk. We describe the algorithm and demonstrate its utility for future genetic studies of whole-genome association.

Natural killer T (NKT) cells express a highly conserved T-cell receptor (TCR) and recognize glycolipids in the context of CD1d molecules. We recently demonstrated that CD4؉ NKT cells are highly susceptible to human immunodeficiency virus type 1 (HIV-1) infection and are selectively depleted in HIV-infected individuals. Here, we identified macaque NKT cells using CD1d tetramers and human V␣24 antibodies. Similar to human NKT cells, ␣-galactosylceramide (␣-GalCer)-pulsed dendritic cells activate and expand macaque NKT cells. Upon restimulation with ␣-GalCer-pulsed CD1d ؉ cells, macaque NKT cells secreted high levels of cytokines, a characteristic of these T cells. Remarkably, the majority of resting and activated macaque NKT cells expressed CD8, and a smaller portion expressed CD4. Macaque NKT cells also expressed the HIV-1/ simian immunodeficiency virus (SIV) coreceptor CCR5, and the CD4 ؉ subset was susceptible to SIV infection. Identification of macaque NKT cells has major implications for delineating the role of these cells in nonhuman primate disease models of HIV as well as other pathological conditions, such as allograft rejection and autoimmunity.Natural killer T (NKT) cells are a subset of T lymphocytes with a highly conserved T-cell receptor (TCR) repertoire in both humans and mice (3). The human NKT cell receptor consists of a V␣24 chain preferentially paired with a V␤11 chain (10,24). While the majority of human NKT cells are CD4 Ϫ CD8 Ϫ , a sizeable portion is CD4 ϩ and a smaller subset is CD8 ϩ (2, 7, 21, 30). NKT cells also display an effector/ memory phenotype based on expression of the memory marker CD45RO and a set of chemokine receptors that is typical of effector T cells (8,18,19,21,22,34).The NKT-cell antigen, albeit elusive, is thought to be presented by an major histocompatibility complex class I-like molecule, CD1d (3). The glycosphingolipid ␣-galactosylceramide (␣-GalCer), which is derived from a marine sponge, is the only known antigen that can bind to CD1d and activate all NKT cells expressing the invariant TCR (31). Activation of NKT cells via their TCR either with anti-TCR antibodies or ␣-GalCer presented by dendritic cells (DCs) results in the rapid secretion of large amounts of cytokines, such as gamma interferon (IFN-␥) and interleukin 4 (IL-4) (6, 9, 25, 27).NKT cells have been implicated in protective immune responses against a wide range of pathogens (5, 15) and in the regulation of autoimmune diseases by suppressing immune responses to autoantigens (13) and by inducing tolerance to antigens exposed in immune-privileged sites (28). Recently we and others have demonstrated that human NKT cells are highly susceptible to human immunodeficiency virus type 1 (HIV-1) infection and are selectively depleted in HIV-infected individuals (21,26,35). Because nonhuman primates are currently the best model system for study of HIV pathogenesis, we sought to identify macaque NKT cells in order to understand their role during HIV infection. Our data revealed the presence of macaque NKT cells whos...

The detection of gene - gene and gene - environment interactions associated with complex human disease or pharmacogenomic endpoints is a difficult challenge for human geneticists. Unlike rare, Mendelian diseases that are associated with a single gene, most common diseases are caused by the non-linear interaction of numerous genetic and environmental variables. The dimensionality involved in the evaluation of combinations of many such variables quickly diminishes the usefulness of traditional, parametric statistical methods. Multifactor dimensionality reduction (MDR) is a novel and powerful statistical tool for detecting and modelling epistasis. MDR is a non-parametric and model-free approach that has been shown to have reasonable power to detect epistasis in both theoretical and empirical studies. MDR has detected interactions in diseases such as sporadic breast cancer, multiple sclerosis and essential hypertension.As this method is more frequently applied, and was gained acceptance in the study of human disease and pharmacogenomics, it is becoming increasingly important that the implementation of the MDR approach is properly understood. As with all statistical methods, MDR is only powerful and useful when implemented correctly. Concerns regarding dataset structure, configuration parameters and the proper execution of permutation testing in reference to a particular dataset and configuration are essential to the method's effectiveness.The detection, characterisation and interpretation of gene - gene and gene - environment interactions are expected to improve the diagnosis, prevention and treatment of common human diseases. MDR can be a powerful tool in reaching these goals when used appropriately.