Extended Data Figure 9: Characterization of DS genes.
From: Enhancer loops appear stable during development and are associated with paused polymerase
a, Defining off-to-on genes. Relative frequency of genes exhibiting a given RPKM value from RNA-seq in whole embryo at 2–4 h and 6–8 h (ref. 9). The threshold (vertical lines) between non-active (off) and active genes was selected based on the local minima in the log-RPKM distribution, as described previously44. DESeq was used to determine if non-active 2–4 h genes had a significant change in their expression at 6–8 h. b, DS-paused genes are expressed at very low levels at 2–4 h, or not at all. Log2 gene expression signal (RPKM) in whole embryo at 2–4 h and 6–8 h (ref. 9) of different categories of genes. Paused DS genes (15 genes, using the stringent criteria for pausing) are significantly less expressed than the top 25% of paused genes (1,776 genes)28 at 2–4 h and are also significantly less expressed at 2–4 h compared to 6–8 h (Mann–Whitney U-test). c, In situ hybridization showing available expression data for DS genes at stage 4–6 (2–4 h) and 11–12 (Berkeley Drosophila Genome Project55). Note, for all 8 genes there is no detectable specific expression at the early time point. d, Log2 GRO-seq expression signal (RPKM) in whole embryo at 2–2.5 h (ref. 28) at the promoter and gene body of different categories of genes. RPKM was defined by Saunders et al.28. e, Histogram (grey bars) of the expected distribution for a gene to be paused (using ‘all’ paused genes defined by Saunders et al.28) by random sampling 20 genes 10,000 times from the 459 differentially expressed off-to-on genes. The red dot indicates the percentage of observed paused genes for all DS genes (differentially expressed but with stable loops). Using this test, the percentage of paused DS genes is highly significant (P 0.0071).
