Fig. 2

Genomic and expression features of HSC lncRNAs. a Classification of lncRNA loci identified in HSC myofibroblasts. Divergent lncRNAs are defined as having a transcription start site (TSS) within 2 kb of the TSS of a protein-coding gene and being transcribed from the strand antisense to the protein-coding gene. Natural antisense lncRNAs overlap the coding gene by at least one base pair. Enhancer-associated lncRNAs are located within 1 kb of an enhancer as defined by H3K27ac. Intergenic lncRNAs have a TSS greater than 2 kb from the TSS of the nearest protein-coding gene and are not contained in any of the other categories. lncRNAs are indicated in red and protein-coding genes are indicated in blue. Arrows show the start and direction of transcription. H3K72ac (K27ac) peaks mark enhancers. The abundance of each class of lncRNA is displayed on the right. b Distribution of the TSS of lncRNAs relative to the TSS of their nearest protein-coding gene. All protein-coding genes were normalized to equal length on the x-axis (black rectangle), and 10 kb of genomic sequence upstream of the TSS of each protein-coding gene is shown. The location of the TSS of each lncRNA was plotted relative to the TSS of the nearest protein-coding gene. lncRNAs that are antisense to protein-coding genes are indicated in red and lncRNAs that are sense to protein-coding genes are indicated in blue. The red peak near the protein-coding TSS indicates that the majority of lncRNAs are located within 2 kb of protein-coding genes and transcribed antisense to protein-coding genes. c lncRNAs (red) are expressed about tenfold lower than their divergent protein-coding genes (blue). The expression levels (log2 transformed reads per fragment per million mapped reads (FPKM)) are indicated on the y-axis. The horizontal black line indicates the mean and open circles indicate outliers. d lncRNAs contain fewer exons than protein-coding genes. The distribution of exons in lncRNA (red) and mRNA (blue) transcripts. e Identification of super-enhancers in HSC myofibroblasts. Enhancers were defined by H3K27ac occupancy and ranked from left to right (x-axis) by the total reads of H3K27ac mapped to each enhancer (H3K27ac signal, y-axis). We identified 321 super-enhancers from a total of 7660 enhancers [70]. f Super-enhancers show broad domains of occupancy compared with typical enhancers. Metagenes represent the mean H3K27ac density (in reads per million unique mapped reads per base pair) across super-enhancers (left) and typical enhancers (right). The metagenes are centered on the enhancer region for each plot and display 3 kb of sequence flanking each enhancer. The median size of a super-enhancer is 18,807 bp and the median size of a typical enhancer is 678 bp. The plots are scaled (x-axis) to reflect the median size of the two classes of enhancers. The increase of signal at the boundary is characteristic of super-enhancers [71] because the boundary represents the H3K27ac peaks at the edges of each super-enhancer. Thus, the peaks at the boundary tend to be aligned with each other while peaks away from the boundaries are distributed more equally. g H3K27ac peaks are enriched at super-enhancers. Metagenes represent the mean H3K27ac density across the major peak of super-enhancers (SE) and typical enhancers (TE). h Example of an lncRNA associated with a super-enhancer. H3K27ac occupancy (normalized logLR value, y-axis) is shown surrounding lncRNA-002221 (red). The domain of the super-enhancer is indicated by a red rectangle. i Super-enhancers were found to be associated with 80 lncRNAs. The median size of super-enhancers associated with lncRNAs was 18,699 bp and the median size of typical enhancers associated with lncRNAs was 1032 bp. We located 968 lncRNAs within 10 kb of a typical enhancer. The enhancers were plotted as described in f. j H3K27ac peaks are enriched at super-enhancers associated with lncRNAs. Metagenes represent H3K27ac density across the major peak of super-enhancers and typical enhancers associated with lncRNAs