Systematic identification of the effect of genetic variation on transcription factor occupancy.

<p>(A) We performed ChIP-seq for the transcription factor CTCF followed by targeted resequencing of its complete occupancy landscape in 12 members of CEPH pedigree 1459 (CEU). (B) Three qualitative levels of occupancy correspond to three genotypes of a SNP located at the binding site, with G/G homozygotes having the highest occupancy (region shown: chr1:151,853,500–151,859,700 [hg18]). (C) The SNP shown in (B) disrupts a critical position in the CTCF consensus sequence (note that G better matches the consensus recognition sequence). (D) Regression of ChIP-seq signal on genotype at the site in (B) quantifies the effect of SNPs on occupancy. We applied this strategy genome-wide to identify sites where SNPs are associated with differences in occupancy. At this site, Akaike information criterion favored a dominant effect model (GT and GG coded identically) over an additive model.