SirenScan reveals the conserved presence of siren RNAs and their evolutionary diversification across angiosperms

Summary Siren RNAs are exceptionally abundant reproductive small interfering RNAs (siRNAs), yet whether they represent a conserved feature of angiosperm reproduction has remained unclear because siren loci lack a standardized definition and identification strategy. We developed SirenScan, a computational framework that integrates cumulation-based and density-based analyses to identify siren loci from small RNA sequencing data. Using SirenScan, we systematically compared ovule and vegetative tissues across representative angiosperm lineages, including the early-diverging angiosperm Nymphaea colorata. We demonstrate that the presence of siren RNAs is a conserved feature of angiosperm ovules, although the degree of expression dominance varies substantially among species. Siren loci are consistently enriched within transposable elements (TEs), while the associated TE superfamilies and cis-regulatory motifs exhibit lineage-specific diversification. In Brassicaceae, siren loci share conserved CLASSY3-associated sequence motifs, whereas other angiosperm lineages display distinct motif compositions, suggesting evolutionary diversification of Polymerase IV targeting mechanisms. Our findings establish siren RNAs as a conserved component of the angiosperm reproductive small RNA landscape and support a model in which conserved RNA-directed DNA methylation machinery is coupled with lineage-specific regulatory mechanisms and transposon landscapes to shape siren locus evolution. SirenScan provides a robust and standardized framework for comparative studies of siren RNA biology and reproductive epigenomics across plant species.