Fig. 3: The strength of CRE loops and microcompartments peaks in AT and weakens as cells enter G1 phase.

a, Asynchronous RCMC contact maps (left) at the Id1, Klf1 and Cdt1 regions with manually annotated interactions shown below the diagonal. Right, zoomed-in boxes shown in greater detail across the M-to-G1 transition using consistent color map scaling. Zoomed-in boxes show examples of microcompartmental CRE loops in i–iv and CTCF/RAD21 loops in v. b, APA plots of loops, separated to show P–P, E–P, E–E, E/P–CTCF and CTCF–CTCF loops across the M-to-G1 transition and for the asynchronous condition. Plots show a 24-kb window centered on the loop at 500-bp resolution and the loops plotted here and in all subsequent panels follow the ‘exclusive’ definition of loop identity the same as in Fig. 2g (CRE sites do not overlap with CTCF). c, Average loop strengths across mitotic exit, with CRE loop strengths on the left axis and CTCF/RAD21-anchored loop strengths on the right axis. Loop strengths were calculated as ‘observed over expected’ signal, the same as in Fig. 2g. d, Change in loop strength across mitotic exit as a function of loop size. The percentage change in total loop strength for each P–P (blue) and CTCF–CTCF (red) loop from AT to LG1 is plotted on the y axis, while the loop size (or interaction distance) is plotted on the x axis. Total loop strengths are calculated as the observed signal without normalization for the expected signal. e, Promoter loop strengths (gray) and RNA Pol II signal (yellow) across mitotic exit, with loop strengths on the left axis and Pol II signal on the right. Loop strengths were calculated as the sum of all observed over expected loop strengths at each promoter and averaged across all promoters. RNA Pol II signal was calculated as the aggregate signal within 1 kb of each promoter-classified loop anchor, averaged across all promoters for each M-to-G1 stage.