Even with sample sizes of hundreds of thousands of individuals, the problem is lack of power to distinguish between true and spurious associations in a GWAS framework. An alternative approach is to integrate independent functional information with GWAS results in order to separate the most likely true associations from spurious signals. The rationale is similar to that of the traditional ‘candidate gene approach’ in principle, where prior functional information is used to effectively limit the number of association tests and thereby to increase power. Indeed, genome-wide functional data have often been used to prioritize among GWAS results and identify the most-promising candidates. For example, the first GWAS for asthma, which resulted in the identification of an associated genomic region containing 19 genes, used gene expression data to hone in on the most promising candidate gene (12). Specifically, gene expression profiles in lymphoblastoid cell lines (LCLs) from a set of asthmatic probands, their parents and siblings, was intersected with the GWAS data, resulting in the implication of ORMDL3 as a novel childhood asthma susceptibility gene (12). In this case, the integration
