To study the location of cis eQTL peaks, we divided each gene into eight bins using the quarter-length of that gene as the unit. For example, bin#1 denotes the region of a gene that is more than 1 unit upstream of the transcription start site (TSS), and bin#2 denotes the region of the gene that is immediately upstream of the TSS. Here we restricted attention to the most highly significant cis eQTLs (those with p ≤ 1E-10) and found the number of peaks (the best hit SNP in the eQTL) in bin#1 through bin#8 were 268, 49, 100, 57, 45, 86, 35, and 251, respectively. A bimodal distribution is apparent in this plot. We surveyed additional p-value cutoffs in addition to 1E-10, and observed consistent results in which the middle part of the genes harbored less cis eQTLs than the TSS and TES regions.
Hao et al. BMC Genetics 2009 10:27 doi:10.1186/1471-2156-10-27