Expanding the horizons of microRNA bioinformatics

Abstract

MicroRNA regulation of key biological and developmental pathways is a rapidly expanding area of research, accompanied by vast amounts of experimental data. This data, however, is not widely available in bioinformatic resources, making it difficult for researchers to find and analyze microRNA-related experimental data and define further research projects. We are addressing this problem by providing two new bioinformatics data sets that contain experimentally verified functional information for mammalian microRNAs involved in cardiovascular-relevant, and other, processes. To date, our resource provides over 4400 Gene Ontology annotations associated with over 500 microRNAs from human, mouse, and rat and over 2400 experimentally validated microRNA:target interactions. We illustrate how this resource can be used to create microRNA-focused interaction networks with a biological context using the known biological role of microRNAs and the mRNAs they regulate, enabling discovery of associations between gene products, biological pathways and, ultimately, diseases. This data will be crucial in advancing the field of microRNA bioinformatics and will establish consistent data sets for reproducible functional analysis of microRNAs across all biological research areas.

Keywords: Gene Ontology; bioinformatic resource; functional annotation; microRNA; network analysis; pathway analysis.

FIGURE 1.

Experimentally verified targets of the miR-29 family. The network was created in Cytoscape (Shannon et al. 2003) using the miRNA:target association data set, “EBI-GOA-miRNA,” available from the PSICQUIC web service. All experimentally validated targets of the three miRNA-29s reported in the literature were curated prior to creating the network. The thin gray edges indicate direct interactions (interaction type = “physical association”); thick gray edges indicate where the directness of the interaction is unknown (interaction type = “association”). The gray nodes are gene targets regulated by the miRNAs; the white nodes are the miRNAs. The RNAcentral identifiers of the miRNAs are as follows: hsa-miR-29a-3p: URS00002F4D78_9606; hsa-miR-29b-3p: URS000024463E_9606; hsa-miR-29c-3p: URS0000272A3D_9606. (*) denotes proteins mentioned in the text.

FIGURE 2.

The role of miR-21 in epithelial-to-mesenchymal transition. (A) Part of the interaction network of hsa-miR-21-5p relevant to epithelial-to-mesenchymal transition (EMT). The network was created using Cytoscape (Shannon et al. 2003) and enriched GO terms identified using the plugins, GOlorize and BinGO (Supplemental Table S3; Maere et al. 2005; Garcia et al. 2007). Enriched GO terms relevant to EMT were selected and the miR-21 targets annotated to these terms or their descendants were clustered together according to the related processes they are involved in. Each entity (node) is color-coded according to the term(s) it is annotated to. Size of the nodes represents the number of times the interaction has been captured as an annotation. The blue edges indicate interaction type “physical association,” applied when the miRNA is demonstrated to bind the mRNA; red edges indicate interaction type “association,” applied when the experimental data does not demonstrate direct miRNA:target binding. (B) The full interaction network of hsa-miR-21-5p; the boxed area shows the target interactions relevant to EMT, as shown in A.