Fig. 2: Assessment of combinatorial TFBS syntax in caQTL enhancers. | Nature Genetics

Fig. 2: Assessment of combinatorial TFBS syntax in caQTL enhancers.

From: Context transcription factors establish cooperative environments and mediate enhancer communication

Fig. 2

a, Distribution of the number of nonoverlapping, above-average top TFBSs across caQTL (n = 4,747, orange or dark shade) and no-effect (n = 4,747, gray or light shade) enhancers. Top, the total number of sites per enhancer. Bottom, the number of sites split by TF class. b, Density of top TFBS positions split by TF class with respect to the ATAC–seq peak center. c, Distribution of average distances of top TFBSs to caQTL (orange) or no-effect (gray) SNPs split by TF class. Initiator TF classes are labeled with a dashed box. Context-only TFs move closer (blue arrow) in caQTL enhancers, while USTs move farther away (red arrow). Differences in distances are assessed using a Mann–Whitney test; P = 0.76, P = 1.1 × 10−10, P = 3.3 × 10−4, P < 2.2 × 10−16 and P = 5.4 × 10−13. d, Assessing combinatorial behavior across TF classes with respect to context-initiator TFs. For each context-initiator TF (TF1, green color), caQTL enhancers are split by whether the SNP creates a new best BS or not. (1) Middle, comparing the top TFBS strength of either unlabeled, UST or context-only TFs (TF2) between the two outlined enhancer groups (Student’s t-test). Shown is the distribution of resulting t-statistics across all possible pairs grouped by the TF2 class. (2) Right, test for the formation of canonical TF heterodimers (TF1–TF2 top TFBS distance <25 bp). Shown are the odds ratios of forming a heterodimer when the caQTL SNP creates a new best BS for the context-initiator TF1 across all possible pairs grouped by the TF2 class. Significance between classes for (1) and (2) is assessed with a Student’s t-test; all P values < 2.2 × 10−16. e, Hierarchical clustering of the t-statistics for the TFBS affinity enrichment test as indicated in d_(1) for individual context-initiator and context-only motif combinations. In c and d, box plots denote the 25th percentile, median and 75th percentile, with whiskers extending to 1.5× the interquartile range. ***P < 0.001, ****P < 0.0001.

Back to article page