Figure 5. Characterization of three categories of intergenic fitCons classes.

a, The density of higher ρ/conserved classes (category I, blue, solid) and lower ρ/open chromatin classes (category II, red, dotted) around protein-coding genes (PCGs) (grey box) suggest a rice promoter of mean size ≈320 bps b, Regions upstream of genes (grey box) with a high fold-change (HFC – purple line) across tissues are enriched for blocks of conserved classes relative to regions upstream of genes with a stable expression across tissues [low fold-change (LFC - dotted line)]. A breakdown by tissue (inset) suggests that regions upstream of genes with differential expression relative to leaf (En=Endosperm; In=Inflorescence; Em=Embryo; Ov=Ovary; Ro=Root and St=Stamen tissues) consistently show greater promoter coverage with blocks of more conserved classes. c, Motif (6–8 bps) enrichment (blue - red) differs between the three categories of noncoding classes (I - conserved, II- open chromatin and III- enhancer candidate classes). Open chromatin classes are relatively enriched for simple repetitive motifs similar to those found in enhancer candidate regions. d, Genes whose upstream regions are enriched for open chromatin classes rather than conserved classes show a broader activation across tissues, but no similar enrichment for differential expression across tissues. Density plots +/− 1 kb around the 1,000 enhancer candidate (EC) sites show e, a defining bi-directional diverged PRO-seq signal (plus strand – blue, minus strand- red, arbitrary strands) identified by dReg⁵⁴, f, a marked enrichment for open chromatin, and g, a generally asymmetric H3K27ac location beyond the nucleosome-depleted core (indicated by the dashed rectangle). EC sites are also associated with h, low ρ, i, low conservation, and, j, a two-fold excess of weak negative selection (Pw).