These studies successfully intersected gene expression profiles with independently obtained GWAS results to provide further support for previously identified candidate loci. Potentially, functional data such as gene expression measurements and GWAS results can also be coupled in a combined analysis. The goal would be to identify novel genetic associations with weak effects, which cannot be distinguished from spurious associations (due to low power) using a standard GWAS approach alone. To date, we know of only few studies that use this paradigm. Naukkarinen et al. (22), for example, collected gene expression data in adipose tissue, and integrated it with results from a GWAS for variation in BMI. Using this approach, they identified 13 nominal genetic associations (P < 0.05) near genes whose expression levels across individuals were correlated with variation in BMI. In turn, Zhong et al. (23) used a similar rationale when they integrated results from an eQTL mapping study with GWAS results for type 2 diabetes (T2D). They found a cis eQTL SNP for the gene ME1, which was also weakly associated with T2D (below genome-wide significance). Subsequent work using a mouse knockout model supported the role of ME1 in determining susceptibility to T2D.
