PRO-seq analysis

PRO-seq data analyses were performed as previously reported⁵⁷. The sequencing reads were aligned to hg38 using Bowtie v.2 using very sensitive parameters. The common artifacts derived from clonal amplification were circumvented by considering maximal three tags from each unique genomic position as determined from the mapping data. To determine E₂-dependent changes in gene body, the sequencing reads for RefSeq genes were counted over the first 13 kb of the entire gene body, excluding the 500 bp promoter-proximal region on the sense strand with respect to the gene orientation by using HOMER⁴. EdgeR⁶⁴ was used to compute the significance of the differential gene expression (fold change (FC) ≥ 1.5, false discovery rate ≤ 0.01). Additionally, a read density threshold (that is, normalized total read counts per kilobase) was used to exclude low-expressed genes. PRO-seqs were normalized to 10 million tags, and HOMER⁴ was used to quantify eRNA expression by tabulating normalized tag numbers surrounding ±1,000 bp from the center of the peaks. eRNAs with a FC > 1.5 in PRO-seq signals were differentially expressed.