Cataloging and comparing long noncoding RNAs generated at enhancers (eRNAs) in prostate, breast, and ovarian cancer cells (744.2)

Abstract

An important purpose of sequencing the human genome was to determine the number of human genes and to catalog them for future study. The final gene count matched the most conservative predictions, with gene sequences representing only 1.5% of the entire human genome. However, more recent transcriptomic analyses have revealed that the cumulative coverage of transcribed regions is at least 75‐85% of the human genome. Moreover, since these analyses did not investigate the process of transcription itself, but only the accumulation of RNAs, the range of cumulative transcribed genome is likely close to the full human genome. The question is how much of the human transcriptome (i.e. the transcribed genome) is functional and how much of the human transcriptome is transcriptional noise. Thus, we return to the question of how many genes (i.e. sequences coding for functional transcripts) are there in the human genome, and to the challenge of cataloging them.In this study, we catalog the full repertoire of long noncoding (lnc)RNAs (length >200 nucleotides) generated at enhancers, also known as enhancer (e)RNAs, in prostate (LNCaP), breast (MCF7), and ovarian (A2780) cancer cells. We use profiles of intergenic DNaseI hypersensitive sites (DNase‐seq) and H3K4me1/3 and H3K27ac regions (ChIP‐seq) to identify enhancers, combined with global run‐on next‐generation sequencing (GRO‐seq) and RNA‐seq to identify, characterize, and compare the eRNA transcriptomes in these three cancer lines. We expect that our analyses will facilitate the study of the functionality of the eRNA transcriptome, and will contribute to distinguishing the fraction of the human transcriptome that is functional from the fraction that could be regarded as noise.