Cell fractionation reveals mis-annotation of human microRNAs

MicroRNAs (miRNAs) form the sequence recognition component of the RNA-induced silencing complex (RISC), bringing Argonaute (AGO) and other RISC-associated proteins to target transcripts where they mediate post-transcriptional gene silencing. High-throughput sequencing has helped to catalogue several thousand human miRNAs, as well as a plethora of other small RNAs derived from larger transcripts. Here, we take a closer inspection of sequencing data derived from cellular fractionation and AGO-RNA immunoprecipitation and find that a subset of small RNAs, despite being denoted as miRNAs, are in fact almost exclusively nuclear and do not associate with AGO. Experimental validation of a select number of these small RNAs also fails to show endogenous miRNA-like activity, suggesting they are degradation products from larger transcripts that have been mis-annotated. Despite this, the presumed miRNA functions of almost 20 of these have collectively been the subject of several hundred publications, emphasising the need to filter unlikely miRNA candidates from sequencing data and large miRNA-sequence repositories. Overall design: miRNA profiles of nuclear and cytoplasmic fractions of HMLE cells without TGF-ß treatment (control) or with TGF-ß (treatment) of 2.5 ng/ml for 2 days or 18 days.