Figure 3.

Characterization of TAD boundaries found in silico. These plots are obtained by analyzing the TAD boundaries found in simulations (through combination of an automated method and visual inspection¹⁹), and by computing the frequencies of non-binding (blue bars), inactive (gray bars) and active (red bars) beads in different sets. Set 1: all beads. Set 2: Beads lying within 100 kbp of a boundary. Sets 3 and 4: The sub-sets of set 2 that also lie within 100 and 20 kbp of a boundary identified in Hi-C data. (i) Beads at boundaries are rich in active and non-binding beads, and depleted of inactive beads (arrows; p values assessed assuming Poisson distributions). (ii) The frequencies of different beads (in sets 1, 2 and 4) in the 150 kbp on each side of either each bead in set 1, or of boundaries in sets 2 and 4. Adapted from Ref.¹⁹