CUT&RUN

CUT&RUN was performed as described (Skene and Henikoff 2017; Skene et al. 2018). We used both anti-GAF (laboratory-made) or anti-NURF301(Novus Biologicals 40360002) at a 1:10 dilution for the antibody binding step. ProteinA-MNase was incubated with calcium on ice for 30 min, and cleaved fragments were recovered by phenol:chloroform extraction. Library prep was performed using the following steps: (1) Ends of digested fragments were repaired by incubation for 30 min at 25°C with 0.5 U/µL T4 PNK, 0.12 U/µL T4 DNA polymerase, and 0.05 U/µL Klenow fragment in 1× T4 DNA ligase buffer (with ATP) and 0.5 mM dNTPs. (2) Fragments were A-tailed by incubation for 30 min at 37°C with 0.25 U/µL Klenow exo- and 0.5 mM dATP in 1× NEBuffer 2. (3) Adapters were added by incubation on the lab bench for 2 h with 4.38 nM laboratory-made Illumina TruSeq forked adapters and 24 U/µL T4 DNA ligase in 1× T4 DNA ligase buffer (with ATP). (4) Library DNA was recovered using AMPure XP beads (1.8× concentration) and PCR-amplified for 15 cycles (all enzymes from New England Biolabs). Libraries were sequenced on an Illumina NextSeq in 37 × 37 paired-end mode. Adapter sequences were removed using fastp (Chen et al. 2018), and reads were aligned to the dm6 reference genome using bowtie2 (Langmead and Salzberg 2012). Only uniquely mapped reads (mapq > 10) with fragment size smaller than 120 bp were retained, and signal coverage tracks were generated using deepTools (Ramírez et al. 2014). Signal was normalized per million mapped reads. See Supplemental Table S5 for alignment metrics.