UCEs comparison to the less constraint SE.

<p>(<b>A</b>) SE are less constraint compared to UCEs. Cumulative distribution plots of phyloP scores of all SE positions (purple line), all UCE positions (red line), random genomic positions (orange line) and SE rare (MAF<0.5%, green line) and prevalent (MAF>5%, black line) SNVs. Prevalent and rare SNVs are extracted from the 1 KG project using global MAFs. Shaded grey area represents 95% confidence interval (obtained by bootstrapping) of random UCE positions (blue line). The numbers of analyzed SNVs are given in the parentheses. (<b>B</b>) SE have a higher overlapping TFBSs count compared to UCEs. Box plots represent results of one hundred sets (each set contains one thousand randomly chosen positions). The y-axis indicates actual ENCODE TFBS overlap per one thousand tested positions. Boxes show IQR, notches indicate 95% confidence intervals of the median, whiskers extend to 1.5 times the IQR and open circles show outliers. *** P <2.2 ×10⁻¹⁶, two- tailed Mann–Whitney test. (<b>C</b>) Venn diagram showing overlap of ENCODE TF and UCE bound TF described by Viturawong et. al <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0110692#pone.0110692-Viturawong1" target="_blank">[12]</a>. (<b>D</b>) Comparison of ENCODE TF and UCE bound TF <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0110692#pone.0110692-Viturawong1" target="_blank">[12]</a> protein domains identifies RNA recognition domain,RRM1 (marked with dashed circle), as the most prevalent domain among UCE bound proteins. Protein domain (Pfam) annotations were done by using the Perseus module in the MaxQuant software suite.