AI is an intuitively advantageous method of detecting cis- regulatory effects. Theoretically, correlating AI with genotypic status should be more sensitive than eQTL based analyses because the comparison is made within an individual (using the other chromosome as a control) thereby removing sources of environmental noise (Fogarty et al., 2010). By applying AI to the TFs and correlating the AI with genotypic status at the risk loci, we demonstrated that MYC and ESR1 are significantly imbalanced in individuals that are heterozygous for the risk locus. A prior report also demonstrated allelic imbalance in MYC expression levels in a colon cancer cell line heterozygous for the 8q24 colon cancer risk allele (Wright et al., 2010). KLF4 did not reveal allelic imbalance most likely due to the low prevalence of informative markers in the KLF4 transcript. The AI method requires appreciable numbers of heterozygous (informative) sites in expressed transcripts. Nevertheless, KLF4 is a strong candidate gene. In prior studies, knockdown of KLF4 in MCF-7 cells affects proliferation, migration, and invasion (Akaogi et al., 2009; Yu et al., 2011). In addition, KLF4 has
