Copy number evolution of conserved noncoding RNA gene families across mammals

Abstract Noncoding RNAs (ncRNAs) play essential roles in mammalian gene regulation, yet their copy number evolution across mammals remains poorly characterised. Here, using high-confidence Covariate Models (CMs) from Rfam, we applied a unified INFERNAL-based framework to systematically predict copy numbers of 588 well-curated, conserved ncRNA gene families across 60 mammalian genomes. Phylogenetic correlation tests indicated that variation in ncRNA family number and total copy numbers is largely independent of genome size, scaffold number, or assembly completeness, indicating that lineage-specific evolutionary processes, rather than genome characteristics, primarily shape ncRNA copy number diversity. Using CAFE analysis, we inferred a burst of ncRNA expansions in the eutherian ancestor, predominantly involving snRNA and H/ACA snoRNA families. In contrast, estimates of phylogenetic signals (Pagel’s λ) and evolutionary rates (σ 2 ) revealed that many miRNA families exhibit lineage-specific copy number expansions. Analysis of highly variable ncRNA families further showed frequent overlap with transposable elements and distinct duplication strategies, including tandem clustering (e.g., mir-154 and SNORA116 ) and dispersed distributions with regional tandem duplications (e.g., 5S_rRNA ). Collectively, these results provide a comparative overview of copy number variation among conserved ncRNA families in mammals. While individual predicted copies may include pseudogenised or TE-derived loci, our analyses reveal robust evolutionary patterns at the family level and provide a foundation for future studies of ncRNA duplication and diversification.