Fig. 3
From: Chromatin activity identifies differential gene regulation across human ancestries
Enhancer differential activity is linked to sequence-dependent differences in NF-\(\upkappa\) B and JunD binding. a Schematic depicting a simplified version of the enrichment test results shown in b. The boxed gradient callout from the enriched overlap shows a hypothetical example of a bQTL where the “G” allele increases the binding affinity of the TF, which leads to elevated mRNA levels through increased transcription of the target gene. b TF bQTL and H3K4me3 QTL enrichments in diff-ATAC, ChIP, and HiC CREs are plotted as Fisher’s exact test odds ratios with error bars representing the lower bound of the 95% confidence interval. Since these are one-sided tests, the upper bound (infinity) is not shown. The total number of CREs used in each test is shown to the right of each odds ratio with asterisks indicating if the P-value passed multiple test correction (*, **, ***, and **** for Bonferroni-corrected P-value < 0.05, 0.005, 5×10−4, and 5×10−5, respectively). c Results of tests on the same CREs used in b separated by their status as enhancer or promoter CREs. d Schematic depicting the directionality matching enrichment test results shown in e. Top, three hypothetical variants with “A” and “G” alleles in two populations where the first (leftmost) has no difference in allele frequency between populations or FST=0, the second (middle) has the maximum possible difference in allele frequency between populations (“A” is fixed in population 1 and “G” is fixed in population 2) or FST=1, and the third (rightmost) has the “G” allele at intermediate frequency in population 2 or FST≈0.11 in this case. Middle, diff-CRE bQTL from overlaps in a–c (gray gradient callout) were tested for enriched matching of ancestry-associated direction. Bottom, schematic of a matching direction diff-CRE bQTL. The blue gradient callout shows an example of a matching direction diff-CRE bQTL where the high affinity “G” allele from the same hypothetical bQTL shown in a is at higher frequency in AFR, which is the ancestry that has the higher ATAC or ChIP scores defining the diff-CRE. e Diff-CRE bQTL matching direction enrichment test results for diff-CREs defined by each displayed score type are plotted as Fisher’s exact test odds ratios with error bars representing the lower bound of the 95% confidence interval. Separate tests were performed on all diff-CRE bQTL (black) and those in the top 5% of FST values among all variants in CREs genome-wide (purple). Since these are one-sided tests, the upper bound (infinity) is not shown. The total number of CREs used in each test is shown to the right of each odds ratio with asterisks indicating if the P-value passed multiple test correction (*, **, ***, and **** for Bonferroni-corrected P-value < 0.05, 0.005, 5×10−4, and 5×10−5, respectively)