Abstract
MicroRNAs (miRNAs) regulate gene expression post-transcriptionally, primarily through binding sites in 3' untranslated regions (3' UTRs). While computational and biochemical approaches have been developed to predict miRNA binding sites on target messenger RNAs, reliable and high-throughput assessment of the regulatory effects of miRNAs on full-length 3' UTRs can still be challenging. Utilizing a miniaturized and high-throughput reporter assay, we present a 'pilot miRNA-targeting map', containing 4,994 successfully measured miRNA:3' UTR regulatory outputs by pairwise assays between 461 miRNAs and eleven 3' UTRs. This collection represents a large experimental miRNA:3' UTR dataset to date on a single platform. The methodology can be generally applied to studies of miRNA-mediated regulation of critical genes. We found that seedless sites can lead to substantial downregulation. We utilized this dataset in the development of a quantitative total score for modeling the total regulatory effects by both seed and seedless sites on a full-length 3' UTR. To assess the predictive value of the total score, we analyzed data from mRNA expression and proteomics studies. We found that the score can discriminate the potent miRNA inhibition from the weak inhibition and is thus useful for quantitative prediction of miRNA regulation. The score has been added to the STarMir program of the Sfold package now available via GitHub at https://github.com/Ding-RNA-Lab/Sfold.