Fig. 1

Overview of the TF profiling model. A Cartoon representing the co-localization between bidirectional transcription observed in nascent RNA sequencing (blue and red are data on each strand) and TF motifs. The PSSM for AP2B is shown. This co-localization can be used to assess global motif displacement scores. B Heatmaps representing the motif displacement distribution [21] for three distinct TFs with different activation states, OFF (ZN586), ON-UP (SP3), and ON-DOWN (PAX5). The center of the heatmap is the position of the middle of the bidirectional (PolII initiation site) and the heat (darker is more) represents the number of motif instances at that position (relative to the center) genome-wide. C Observed promoter (top) and non-promoter (e.g., enhancers, bottom) per position base probabilities surrounding PolII initiation sites show a profound GC bias. In this data set [38], bidirectionals are 30% at promoters (top) and 70% at non-promoters (enhancers, bottom). D The observed motif displacement score distribution assuming a flat background and no positional information (left) compared to a position dependent di-nucleotide Markov background (right). Each dot is a single TF position specific scoring matrix, colored by its inherent GC content. The probability (\(p_i\)) is defined by the observed probabilities (N) at position i. The position and motif displacement distribution for AP2B is shown with both background models