Binding of ELF2 impacts AFE shift following selenium treatment.

<p>A) Model illustrating the situation where transcription is shifted towards the downstream AFE following selenium treatment and demonstrates an example of a TSS that would be included in our analysis. B) Chromatin accessibility of TSS with significant shift in AFE following selenium measured by ATAC-seq (read count normalized to total reads in the library). Chromatin accessibility profiles derived from LCLs treated with selenium are bright blue or red, while profiles derived from LCLs treated with control are dark blue or red. Blue lines show accessibility at TSS to which the AFE shifts towards in selenium while red show the other TSS (as illustrated in A). C and D) ATAC-seq profiles centered on ELF2 motif locations within 1000bp of either TSS (colors are the same as in A and B) where C) shows accessibility of ELF2 motifs near TSS that are favored following selenium and D) shows accessibility of ELF2 motifs near TSS that are favored in the control samples. The difference in the ratio of treatment vs control read counts between the preferred and not preferred AFE is significant (BH p-value = 2.0 × 10<sup>−3</sup>). E) QQ-plot of the AFE QTL p-values for SNPs in footprint of ELF2 (within cis region of an AFE event) that are predicted to influence binding (black) or not (gray). F) Association between genotype of SNP (rs546382) found in an ELF2 footprint and predicted to influence binding [<a href="http://www.plosgenetics.org/article/info:doi/10.1371/journal.pgen.1006995#pgen.1006995.ref048" target="_blank">48</a>] and Ψ of AFE in <i>IGHMBP2</i> across European individuals from the GEUVADIS data (<i>p</i>-value = 1.1 × 10<sup>−35</sup>). In the bottom right of the graph is the motif logo for ELF2 and the arrow indicates the position of the SNP.</p>