Intrinsically linked lineage-specificity of transposable elements and lncRNAs reshapes transcriptional regulation species- and tissue-specifically

Abstract

AbstractLineage-specificity of transcriptional regulation by lncRNAs critically determines whether mouse models reliably mimic human diseases. To address this question, we identified human/mouse-specific lncRNAs from GENCODE-annotated human and mouse lncRNAs, predicted their DNA binding domains (DBDs) and binding sites (DBSs), analysed transposable elements (TEs) in DBDs and DBSs, and analysed functional enrichment of target genes. 84%/98% of human/mouse-specific lncRNAs, 61%/95% of their DBDs, and 46%/73% of their DBSs contain TEs almost exclusively originated from simians/rodents, indicating intrinsically linked lineage-specificity of TEs, lncRNAs, and lncRNAs’ DBSs. We then revealed how transcriptional regulation is lineage-specifically rewired by co-opted lncRNAs and DBSs by analysing distributions of target genes in signalling pathways and expression of target genes in multiple tissues in humans and mice. Transcriptional regulation is greatly rewired species-specifically and tissue-specifically. We further analysed transcriptomic data of Alzheimer’s disease and tumours from human patients and mouse models, with results supporting the above conclusions. Our results reveal the intrinsically linked lineage-specificity of transposable elements, lncRNAs, and transcriptional regulation, provide data and tool for analysing and differentiating transcriptional regulation in humans and mice, and suggest that many evolutionary novelties may be destined to be lineage-specific.