gms | German Medical Science

A leading hypothesis for the mechanism by which air pollution causes asthma and other respiratory effects is through oxidative stress, a complex pathway involving a host of genes for the metabolic activation and detoxification of oxidizying agents [Ref. 1]. I will review the epidemiologic evidence from the Southern California Childrens Health Study for the modifying effects of family history, various candidate genes, and other host susceptibility factors on ozone, particulates and other pollutants, and show how these could be organized in a unified statistical model. This motivates the development of new statistical approaches to looking a gene-environment and gene-gene interactions with multiple genes and multiple environmental or host factors in combination, including empirical approaches combining hierarchical models with Bayesian model averaging techniques and Bayesian pharmacokinetic modeling of the entire pathway [Ref. 2]. Recurring themes involve allowing for uncertainty about model form in estimating model parameters and incorporation of prior knowledge in the form of covariates for some parts of the model. While traditional linkage analysis and candidate gene studies have identified a number of genes or chrosomal regions that may be involved, there probably remain many still to be discovered. Recent genotyping technology advances have now made it possible to consider genome-wide association scans involving, say, 600,000 SNPs on large case-control samples. Cost-efficient study designs and analysis methods that would extract the maximum possible information out of such studies are an active area of research [Ref. 3].