Fig. 6: Context-only TFBSs are linked to high coactivator binding.
From: Context transcription factors establish cooperative environments and mediate enhancer communication
a, Fraction of 20-amino-acid windows with low complexity (y axis) within the non-DBD of TFs (schematics) across TF classes. Significance is computed using unlabeled TFs as a reference or by comparing across classes (Mann–Whitney test, comparison to unlabeled class: P = 0.97, P = 0.36, P = 0.06 and P = 2.5 × 10−3, respectively, PUST versus Pcontext-only = 2.9 × 10−3, Pcontext-initiator versus Pcontext-only = 2.2 × 10−4). Box plots denote the 25th percentile, median and 75th percentile, with whiskers extending to 1.5× the interquartile range. b, Schematic representation explaining the model used in c and d. Accessibility drives baseline levels of H3K27ac and BRD4 immunoprecipitation coverage (red shaded area represents the expected enrichment). TFs are evaluated on whether their top BS strength is associated with more or less immunoprecipitation coverage after controlling for accessibility. c,d, Results of a log-linear model associating immunoprecipitation coverage (c, H3K27ac and d, BRD4) with top TFBS strength within an enhancer while using ATAC–seq coverage as a covariate. The t-statistics of the corresponding BS strength model coefficients are plotted on the y axis. TFs are split by TF class (a), and significance is assessed with respect to unlabeled TFs (Mann–Whitney test; P = 1.2 × 10−14, P = 0.52, 6 × 10−7 and P = 3.1 × 10−14 for c and P < 2.2 × 10−16, P = 1.1 × 10−3, P = 0.38 and P < 2.2 × 10−16 for d). Violin plots encompass the entire data range and represent the medians. **P < 0.01, ***P < 0.001 and ****P < 0.0001.
