Well-annotated microRNAomes do not evidence pervasive miRNA loss

microRNAs are conserved non-coding regulatory factors implicated in diverse physiological and developmental processes in multicellular organisms, as causal macroevolutionary agents and for phylogeny inference. However, the conservation and phylogenetic utility of microRNAs has been questioned on evidence of pervasive loss. Here we show that apparent widespread losses are, largely, an artefact of poorly sampled and annotated microRNAomes. Using a curated dataset of animal microRNAomes, we reject the view that miRNA families are never lost, but they are rarely lost (92% are never lost). A small number of families account for a majority of losses (1.7% of families account for >45% losses), and losses are associated with lineages exhibiting phenotypic simplification. Phylogenetic analyses based on the presence/absence of microRNA families among animal lineages, and based on microRNA sequences among Osteichthyes, demonstrate the power of these small datasets in phylogenetic inference. Perceptions of widespread evolutionary losses of microRNA families are due to the uncritical use of public archives corrupted by spurious microRNA annotations, and failure to accommodate false absences that occur because of incomplete microRNAome annotation.