miRDB: an online resource for microRNA target prediction and functional annotations.
Bottom Line: Although thousands of miRNAs have been identified, many of the reported miRNAs are not likely to play active functional roles or may even have been falsely identified as miRNAs from high-throughput studies.To address this issue, we have performed combined computational analyses and literature mining, and identified 568 and 452 functional miRNAs in humans and mice, respectively.These miRNAs, as well as associated functional annotations, are presented in the FuncMir Collection in miRDB.
Affiliation: Department of Biomedical Engineering, Washington University, St. Louis, MO 63130, USA Department of Radiation Oncology, Washington University School of Medicine, St. Louis, MO 63108, USA.Show MeSH
Mentions: A composite score was calculated for each miRNA by summarizing four individual criterion scores described above (Figure 3A). A score of 3 or higher was used to define functional miRNAs. In this way, 568 and 452 functional precursor miRNAs were identified for humans and mice, respectively, which represent about one-third of all miRBase entries for the two species. Interestingly, the four individual criteria identified largely overlapping miRNAs. Only 1% of all functional miRNAs were identified by a single criterion only (PubMed Score), whereas two-thirds were identified by at least three of the four criteria (Figure 3B). The relative contribution of each criterion was also evaluated by removing the corresponding individual criterion score from composite score calculation. PubMed Score had the largest contribution, and was required for the identification of 183 (32%) functional miRNAs in humans. Similarly, Expression Score, miRBase Score and Conservation Score were required for the identification of 165, 29 and 20 functional human miRNAs, respectively. The complete list of both composite scores and individual criterion scores for all the miRNAs is presented in Supplementary File 1. For each functional precursor miRNA, the dominantly expressed mature miRNA was selected. In some cases, a second mature miRNA was selected from the same precursor if its expression level was no less than 10% of that of the dominant mature miRNA and its read count was no less than 5 reads per million reads as determined by RNA-seq experiments. In this way, 654 and 442 functional mature miRNAs were identified in humans and mice, respectively. All these identified functional miRNAs were included in the FuncMir Collection of miRDB, with one entry homepage for each miRNA. A screenshot of a representative miRNA entry is presented in Figure 3C. The CLASH RNA-seq data (11) were analyzed to evaluate the expression profiles of FuncMir miRNAs. Among all 2588 human mature miRNAs included in miRBase, 322 were detected at least five times per million sequence reads in HEK293 cells. Most of these expressed miRNAs (79.2%) have been included in the FuncMir Collection. In summary, the FuncMir Collection helps to facilitate researchers to focus on functionally relevant miRNAs, especially in experimental screens for miRNA functions or systematic computational analysis at the genome level.
Affiliation: Department of Biomedical Engineering, Washington University, St. Louis, MO 63130, USA Department of Radiation Oncology, Washington University School of Medicine, St. Louis, MO 63108, USA.