BACKGROUND: Numerous studies over the past decade have elucidated a large set of long intergenic noncoding RNAs (lincRNAs) in the human genome. Research since has shown that lincRNAs constitute an important layer of genome regulation across a wide spectrum of species. However, the factors governing their evolution and origins remain relatively unexplored. One possible factor driving lincRNA evolution and biological function is transposable element (TE) insertions. Here, we comprehensively characterize the TE content of lincRNAs relative to genomic averages and protein coding transcripts.
RESULTS: Our analysis of the TE composition of 9,241 human lincRNAs revealed that, in sharp contrast to protein coding genes, 83% of lincRNAs contain a TE, and TEs comprise 42% of lincRNA sequence. lincRNA TE composition varies significantly from genomic averages - L1 and Alu elements are depleted and broad classes of endogenous retroviruses are enriched. TEs occur in biased positions and orientations within lincRNAs, particularly at their transcription start sites, suggesting a role in lincRNA transcriptional regulation. Accordingly, we observed a dramatic example of HERVH transcriptional regulatory signals correlating strongly with stem cell-specific expression of lincRNAs. Conversely, lincRNAs devoid of TEs are expressed at greater levels than lincRNAs with TEs in all tissues and cell lines, particularly in the testis.
CONCLUSIONS: TEs pervade lincRNAs, dividing them into classes, and may have shaped lincRNA evolution and function by conferring tissue-specific expression from extant transcriptional regulatory signals.