Intrinsically linked lineage specificity of transposable elements, lncRNA genes, and transcriptional regulation

Abstract

Unlike protein-coding genes, many mammalian lncRNAs are lineage-specific (LS), but which human and mouse reference lncRNAs are lineage-specific, how they are generated, and whether their regulatory elements are lineage-specific remain unclear. To resolve their origins, properties, and impacts, we systematically classified primate-, simian-, rodent-, human-, and mouse-specific lncRNAs in GENCODE-annotated human and mouse genomes using structure-aware covariance modeling. We predicted their DNA-binding domains (DBDs) and genomic binding sites (DBSs) and traced the origins of embedded transposable elements (TEs). LS TEs contribute to over 84% (human) and 98% (mouse) of LS lncRNAs, 61% and 95% of their DBDs, and 46% and 73% of their DBSs. Analyses of multi-omics data and a developmental process suggest that LS TE-linked lineage-specificity substantially rewires transcriptional regulation and that many evolutionary novelties may be structurally predisposed. These findings bridge the gap between human diseases and mouse models, providing valuable resources for identifying human-specific therapeutic lncRNA targets.