Abstract
Modern microarray genotyping now permits simultaneous analysis of tens of thousands of polymorphisms, and this technology is being widely used to associate the role of genes with the etiology of complex disease. Genome-wide hypothesis-free mapping will also increasingly generate candidate genes that require further testing in association studies. At the same time, genetic effects are increasingly observed to be buffered by a wide array of biologic mechanisms that evolved to protect the genome from environmental insult and that serve to obscure observation of direct effects of polymorphisms on a disease phenotype. These two forces combine to make replication of genomic epidemiology extraordinarily difficult. Traditional research synthesis of emerging bodies of genomic epidemiology is problematic and often quickly outdated. The author proposes that electronic evidence-based methodology, perhaps modeled after that used by the Cochrane Collaboration in clinical medicine, would facilitate the systematic preparation and frequent updating of systematic reviews, which is essential for identifying valid and replicable gene-disease associations.