(Molecular Cell 75, 284–297.e1–e6; July 25, 2019) In the original online version of this paper, a subset of panels in Figures 5, S5, and 6 were generated with a consensus peak set that erroneously omitted 5,501 peaks (original, n = 133,656; corrected, n = 139,157). This omission originated from an undetected run error in our peak calling script for the 1 h mature repli-ATAC-seq datasets. This minorly changed the numbers of peaks in each clustering category (Figures 5A and 5B) but had no influence on the trends we observed (Figures 5C–5G, Figures S5A–S5D, Figures S5F–S5H). It similarly had no effect on the behavior of our clustering categories in transcription-inhibited repli-ATAC-seq (Figure 6E). Enriched motifs with low p values in each clustering category changed in the corrected data (Figure S5C). The figures and associated text have been corrected and now appear with the paper online and in print. These corrections do not change the conclusions of the study. The authors regret this error. [Figure presented] [Figure presented] [Figure presented] [Figure presented] [Figure presented] [Figure presented]