Too Many False Targets for MicroRNAs: Challenges and Pitfalls in Prediction of miRNA Targets and Their Gene Ontology in Model and Non-model Organisms

Short (“seed”) or extended base pairing between microRNAs (miRNAs) and their target RNAs enables post-transcriptional silencing in many organisms. These interactions allow the computational prediction of potential targets. In model organisms, predicted targets are frequently validated experimentally; hence meaningful miRNA-regulated processes are reported. However, in non-models, these reports mostly rely on computational prediction alone. Many times, further bioinformatic analyses such as Gene Ontology (GO) enrichment are based on these in silico projections. Here such approaches are reviewed, their caveats are highlighted and the ease of picking false targets from predicted lists is demonstrated. Discoveries that shed new light on how miRNAs evolved to regulate targets in various phyletic groups are discussed, in addition to the pitfalls of target identification in non-model organisms. The goal is to prevent the misuse of bioinformatic tools, as they cannot bypass the biological understanding of miRNA–target regulation.