Fig. 2. Cell-state-specific unannotated RNAs make up a large fraction of the caRNAs.
a Schematic of the method used to catalog unannotated RNAs by identifying transcription units using StringTie2. b Genome tracks showing the chromatin context of 3 representative unannotated transcription loci (UTL). Left panel: UTL69162 and UTL69163, respectively, downstream and antisense to RB1CC1, are classified as readthrough RNA and CRE-derived RNAs. Right panel: UTL69657 is classified as a repeat-derived RNA due to its overlap with a LINE element. In both left and right panels, the top 2 tracks display the strand-specific genome coverage of the RNA-derived side of the ChAR-seq reads in ES and DE replicate 1 (+ strand ES in dark blue, − strand ES in light blue, + strand DE in dark yellow, − strand ES in light yellow). The next two tracks display the strand-specific genome coverage of the total RNA-seq data. c Relative composition of the chromatin-associated UTLs in the 7 annotation classes. d Scatter plots showing the chromatin association scores for individual UTLs and their abundance in the caRNA population. Chromatin-enriched and depleted UTLs were determined using DESeq2 (FDR 0.05, fold change threshold 3x). Pie charts summarize the fraction of chromatin-enriched and chromatin-depleted UTLs in each category. Numbers within each pie chart indicate the total number of RNAs in that category. e Percentage of genes upregulated and downregulated in DE vs ES cells in the caRNA transcriptome and for each RNA category. Up- and downregulated RNAs were identified using DESeq2 (FDR 0.05, fold change threshold 3x). f RNA-DNA contact maps in ES and DE cells for the top 200 most abundant UTLs on Chr7 and Chr8, displayed at a resolution of 1 RNA per row and 1 Mbp of genome space per column. g Genome-scale chromatin interaction profiles of 4 UTLs showing similar localization patterns as annotated RNAs.