Computational Biology Group at the Faculty of Mathematics,
Informatics and Mechanics, University of Warsaw

Comprehensive prediction in 78 human cell lines reveals rigidity and compactness of transcription factor dimers

Aleksander Jankowski, Ewa Szczurek, Ralf Jauch, Jerzy Tiuryn and Shyam Prabhakar

Supplemental figures and tables submitted with the manuscript:

  • Supplemental Figure 1. Cell type dendrogram. 78 ENCODE cell types were hierarchically clustered by the degree of overlap of their hypersensitive regions, resulting in 41 cell type clusters (see Methods). The inner structure of the clusters is shown in black, whereas the relationships between the clusters are shown in red.
  • Supplemental Figure 2. The effect of motif complex orientation. As in Fig. 1, but indicating the number of motif complex occurrences for both of the possible motif orientations in a motif complex.
  • Supplemental Figure 3. Cluster of highly similar motif complexes corresponding to AR–FOXA1 cooperativity. The top 10 overrepresented motif complexes comprising row 6 in Fig. 2 are shown. The color of the motif bounding boxes indicates motif orientation relative to TRANSFAC: blue – same orientation, red – opposite.
  • Supplemental Figure 4. Same as Fig. 5, but after motif trimming (see Methods).
  • Supplemental Figure 5. Converging FOXA1 homodimer 3D structure (Fig. 6C) seen from a different perspective. Notably, C termini do not overlap.
  • Supplemental Table 1. Known dimeric DNA-binding transcription factor complexes.
  • Supplemental Table 2. DNase I hypersensitivity datasets used in this study. Datasets covering 78 distinct cell types, generated by the ENCODE group at the University of Washington, were obtained from the UCSC Genome Browser (wgEncodeUwDnase track, hg19 read alignments).
  • Supplemental Table 3. Cell-type–specific statistics of our predictions. For each of the 41 cell types (i.e. cell type clusters) we indicate the number and total length of cell-type– specific hypersensitive sites, the number of overrepresented motif complexes and the number of predicted cooperative interactions.
  • Supplemental Table 4. Detailed list of all 603 predicted cooperative interactions. Each of the 5,233 rows corresponds to an overrepresented motif complex. Their clustering into 603 predictions is indicated in the cluster_id column, with the first row in each cluster corresponding to signature motif complex.

In addition to the above, the following supplemental datasets are available: