Combining molecular and genetic data from different sources

The rapidly growing number of molecular epidemiology studies is providing an enormous, often multidimensional, body of evidence on the association of various disease outcomes and biomarkers. The testing and validation of statistical hypotheses in genetic and molecular epidemiology presents a major challenge requiring methodological rigor and analytical power. The non-replication of many genetic and other biomarker association studies suggests that there may be an abundance of spurious findings in the field. This chapter will discuss ways of combining evidence from different sources using meta-analysis methods. Research synthesis not only aims at producing a summary effect estimate for a specific biomarker, but also offers a unique opportunity for a meticulous attempt to critically appraise a research field, identify substantial differences between or within studies, and detect sources of bias. Systematic reviews and meta-analyses in human genome epidemiology are specifically discussed, as they comprise the bulk of the available evidence in molecular epidemiology where these methods have been applied to date. Considered here are issues regarding validity and interpretation in genetic association studies, as well as strategies for developing and integrating evidence through international consortia. Finally, there is a brief look at how combining data through meta-analysis may be applied in other areas of molecular epidemiology.