Development of Bioinformatics Software Tools for Network- and Pathway-Based Exploration of miRNA Target Interactions

miRNAs are short, non-coding RNA involved in the post-transcriptional regulation of mRNA in both health and disease. A single miRNA can target hundreds or even thousands of mRNAs. Similarly, a single mRNA can be targeted by multiple miRNAs simultaneously. Moreover, the functional impact of miRNA targeting can be difficult to elucidate due to overlapping and redundant targeting, heterogenous tissue expression, direct and indirect effects, and convergent signalling mechanisms.

The large interest in miRNA research and miRNA-target interactions (MTIs) has led to the development of many publicly available MTI databases (DBs) and other tools. The use of individual DBs and tools, however, is error-prone, due to rapidly evolving miRNA nomenclature guidelines and resultant naming inconsistencies within publications and other resources, similar naming inconsistencies across molecular identifiers and gene symbols, and the use of deprecated resources. MTI prediction algorithms also suffer from a high level of false positives.

A standard workflow for MTI investigation can include miRNA target identification, functional annotation of miRNA targets, computation of overlapping targets, and pathway enrichment analysis, all of which require application of individual tools and software packages. Further errors, however, caused by ID mismatches and data inconsistencies within and across combined resources, can be introduced at many points in these workflows, leading to irreproducible research and potentially incorrect conclusions. At worst, these issues have directly contributed to the retraction of publications.

To address these challenges, we here developed and applied three software tools to perform optimised miRNA and MTI research: 1) miRNAmeConverter, an R-package and web application that performs high-throughput miRNA name translation. 2) GeneNameGenie, a molecular ID graph database and translation framework that maps relationships between 47 molecular ID types and facilitates rapid translation of common IDs, including miRNAs. GeneNameGenie also performs integrated Reactome pathway enrichment analysis. 3) miRGIK builds on the GeneNameGenie graph database and enriches it with more than 48 million MTIs and their metadata, gene-tissue expression data, and transcription factor (TF) target information. miRGIK is a complete, standalone MTI graph database resource and research tool which enables sophisticated MTI identification and filtering including tissue expression, top-down (pathway - gene - miRNA) and bottom-up (miRNA - gene - pathway) functional analysis of miRNA direct and indirect targeting, computation of overlapping and redundant targets, detailed TF targeting investigation, and pathway enrichment analysis.

The developed software tools can complement wet-lab research to facilitate holistic investigation of miRNA functions in a biological context, and were herein applied to analyse the functional role of miRNAs dysregulated in ischaemic neuronal injury (Pfeiffer et al., 2021) and epilepsy (Venø et al., 2020).