Fig. 5: TAD dynamics is consistent with uniform cohesin dynamics and is governed by CTCF binding.
From: Uniform dynamics of cohesin-mediated loop extrusion in living human cells
a, Deviation of simulations from experiments, assuming a cohesin motor speed of 1 kb s−1 (top) or 0.25 kb s−1 (bottom). Black squares correspond to nonassessed parameter combinations. b, Deviation of simulations from experiments, separately considering Micro-C (black) or live-cell imaging (colors) data, and assuming a cohesin motor speed of 1 kb s−1. c, Superposed contour plots for all genomic regions, based on Micro-C and live-cell imaging data taken together, for a motor speed of 1 kb s−1 (left) or 0.25 kb s−1 (right). In a–c, solid and dashed lines indicate the 10% and 25% best parameter sets, respectively. d,e, Average squared 3D anchor–anchor distance time series weighted by localization precision, aligned on the starting times of proximal states (\({t}_{{\rm{start}}}\)), from experiments (d) and simulations (e) of the L1 TAD. The shaded area indicates the weighted s.e.m. Simulations (e) assume cohesin motor speeds ranging from 0.125 kb s−1 (top) to 1 kb s−1 (bottom), as indicated, and assume a cohesin density and residence time of 12 Mb−1 and 22 min, respectively. Dashed lines show a fitted piecewise linear function used to estimate the CR. For simulations with motor speeds of 0.5 and 1 kb s−1, the time series do not exhibit a linear decrease and CR cannot be estimated. Panels d and e were created in part with BioRender.com.
