Detailed characterization of gene loops. (A) Chromatin loop sizes and direction around TSSs grouped according to gene expression level. (B) Chromatin loop sizes and direction from various anchor points around TSSs, as a function of gene length. (C) Identification of genes having self-looped structures. For each gene, the search radius “R” was based on the distance between the focal point (red dot) and TTS. With this strategy, genes with chromatin loops are classified as “3′ only,” “5′ only,” or “both,” depending on the locations of the interacting partners. (D) Fraction of genes with self-looped structures classified as “3′ only” in C. (E,F) Patterns of chromatin loops related to regions upstream of genes. For a focal point located upstream of a TSS, its corresponding chromatin segment can have interacting partners located at 3′ or 5′ regions with respect to the gene (E). A series of focal points (5, 2, 1, 0.5, and 0 kb upstream of TSSs) were analyzed on all genes in the genome divided by expression levels (F). (G) Correlation between gene loop and expression, given as the natural base logarithm of RPKM (reads per kilobase per million reads). Two pairs of comparisons on transcriptional activities are shown, depending on whether there is a highly expressed gene (level 7–9) located within 5 kb of genes of interest. The P-values indicate Wilcoxon-Mann-Whitney test results.