Abstract
microRNAs (miRNAs) are central post-transcriptional regulators that shape cellular functions, development, and phenotypic diversity. Yet their evolutionary history in mammals remains poorly resolved, largely due to incomplete genome-wide annotation and the difficulties in establishing miRNA orthology. Using a homology-based framework paired with genome synteny for orthology assignment, we characterised miRNA families and investigated their evolutionary dynamics across mammals. Compared to Infernal and MirMachine using a consensus pseudo-ground-truth framework, our method demonstrated better sensitivity-specificity balance, allowing divergent miRNA detection while reducing false positives. Analysis of 73 mammalian genomes revealed major miRNA family gains at the ancestral nodes of Eutheria, Marsupialia, and Haplorrhini, contrasted with substantial losses in Eulipotyphla and Scandentia. Copy number evolution was largely conserved across families, with only a few families (e.g., MIR-145, MIR-2285) showing lineage-specific duplication acceleration. Comparative genome microsynteny further resolved orthology for the multi-copy family LET-7, spanning eight deeply conserved syntenic blocks, as well as for 143 individual miRNAs. We showed that most miRNAs are deeply conserved, with the seed region being most constrained, followed by the mature, star, and loop regions. Within the seed region, positions 4-6 displayed higher conservation than other seed positions, consistent with their roles in core target recognition. Ancestral seed-state reconstruction further revealed diverse evolutionary signatures, including ancestral shifts, recurrent substitutions, and homoplasious changes. Collectively, these findings provide an unprecedented comparative overview of mammalian miRNA evolution and create new opportunities to investigate how shifts in miRNA repertoires and sequence features may underlie lineage-specific regulatory and phenotypic diversification.
Full text
2,059 characters
· extracted from
oa-doi-fallback
· click to expand
Abstract
microRNAs (miRNAs) are central post-transcriptional regulators that shape cellular functions, development, and phenotypic diversity. Yet their evolutionary history in mammals remains poorly resolved, largely due to incomplete genome-wide annotation and the difficulties in establishing miRNA orthology. Using a homology-based framework paired with genome synteny for orthology assignment, we characterised miRNA families and investigated their evolutionary dynamics across mammals. Compared to Infernal and MirMachine using a consensus pseudo-ground-truth framework, our method demonstrated better sensitivity-specificity balance, allowing divergent miRNA detection while reducing false positives. Analysis of 73 mammalian genomes revealed major miRNA family gains at the ancestral nodes of Eutheria, Marsupialia, and Haplorrhini, contrasted with substantial losses in Eulipotyphla and Scandentia. Copy number evolution was largely conserved across families, with only a few families (e.g., MIR-145, MIR-2285) showing lineage-specific duplication acceleration. Comparative genome microsynteny further resolved orthology for the multi-copy family LET-7, spanning eight deeply conserved syntenic blocks, as well as for 143 individual miRNAs. We showed that most miRNAs are deeply conserved, with the seed region being most constrained, followed by the mature, star, and loop regions. Within the seed region, positions 4-6 displayed higher conservation than other seed positions, consistent with their roles in core target recognition. Ancestral seed-state reconstruction further revealed diverse evolutionary signatures, including ancestral shifts, recurrent substitutions, and homoplasious changes. Collectively, these findings provide an unprecedented comparative overview of mammalian miRNA evolution and create new opportunities to investigate how shifts in miRNA repertoires and sequence features may underlie lineage-specific regulatory and phenotypic diversification.
Competing Interest Statement
The authors have declared no competing interest.
Text is read by the "Ask this paper" AI Q&A widget below.
Extraction quality varies by source — PMC NXML preserves structure
cleanly, OA-HTML may include some navigation residue, and OA-PDF can
have broken hyphenation. The publisher copy
(via DOI)
is the canonical version.