We next hypothesized that the 15 GWAS-identified risk loci are cis-acting eQTLs of transcription factors (TFs), which in turn influence multiple downstream targets. In such cases, the activity level of a given TF may be more accurately reflected in the expression levels of its target genes (Essaghir et al., 2010; Wolfer et al., 2010). It is then possible to examine the set of eQTL-associated genes and then work “backwards” to evaluate if a TF binding motif is enriched in the target genes (Figure 2). Thus, each risk locus will be associated with a set of target genes. For each set of target genes, the putative enhancer regions, as defined by ENCODE generated DNaseI hypersensitivity (DHS) data from the MCF-7 cell line (a breast cancer cell line), were analysed for TF DNA binding motif enrichment (Methods). This analysis revealed 3 risk loci for which the target genes are significantly enriched for 3 particular TF motifs (6q25/ESR1; 9q31/KLF4, 8q24/MYC) (Table 1). Moreover, as mandated by our analysis, the particular TF whose motif is over-represented in the target genes is physically located nearby
