miRspongeR: an R/Bioconductor package for the identification and analysis of miRNA sponge interaction networks and modules

Junpeng Zhang, Lin Liu, Taosheng Xu, Yong Xie, Chunwen Zhao, Jiuyong Li, Thuc Duy Le

Introduction

MicroRNAs (miRNAs) are small non-coding RNAs that play a pivotal role in gene regulation by typically silencing target mRNAs. However, the regulatory landscape is complex; various RNA molecules—including long non-coding RNAs (lncRNAs), circular RNAs (circRNAs), and even other mRNAs—can act as "miRNA sponges" or competing endogenous RNAs (ceRNAs). These sponges contain miRNA response elements (MREs) that allow them to sequester miRNAs, thereby preventing the miRNAs from binding to their primary mRNA targets and effectively upregulating gene expression.

Despite their importance in diseases like cancer, identifying miRNA sponges is computationally intensive. Previous tools were often fragmented, limited to single algorithms, or restricted to web-based interfaces that hindered large-scale local analysis. miRspongeR addresses these gaps by providing a unified, comprehensive R/Bioconductor framework for the identification and downstream analysis of miRNA sponge interactions.

Methods: An End-to-End Pipeline

The miRspongeR package streamlines the research process through a three-stage pipeline: identification, module discovery, and validation/analysis.

The miRspongeR workflow, illustrating the transition from input data to validated biological insights.
  1. Identification: The package integrates seven established algorithms: miRHomology, pc, sppc, hermes, ppc, muTaME, and cernia. These methods range from "static" approaches based on sequence complementarity to "dynamic" approaches that utilize expression correlation. A key feature is the integrateMethod, an ensemble approach that uses majority voting to identify high-confidence interactions across multiple algorithms.
  2. Module Discovery: To move beyond individual interactions, miRspongeR identifies functional "modules"—clusters of sponges that co-regulate targets—using four network clustering algorithms: FN, MCL, LINKCOMM, and MCODE.
  3. Validation and Analysis: The tool automates the comparison of results against ground-truth databases. It also supports functional enrichment analysis (e.g., GO, KEGG, Reactome) and clinical survival analysis using Cox proportional hazards models to evaluate the prognostic value of identified modules.

Datasets and Case Study

To demonstrate the package's utility, the authors analyzed a Human Breast Invasive Carcinoma (BRCA) dataset from The Cancer Genome Atlas (TCGA), involving matched miRNA and mRNA expression data for 72 individuals. Experimentally validated miRNA-target interactions were sourced from miRTarBase v7.0, and validation was performed against curated databases like miRSponge, LncCeRBase, and LncACTdb v2.0.

Key Insights

  • Ensemble Robustness: The study emphasizes that no single algorithm is definitive. The integrateMethod allows researchers to mitigate the noise of individual computational predictions by focusing on interactions supported by multiple lines of evidence.
  • Dynamic Context: While sequence homology provides a baseline, incorporating gene expression data through dynamic methods is crucial for understanding sponge behavior in specific disease states.
  • Clinical Translation: By integrating survival analysis, miRspongeR bridges the gap between network biology and clinical application, identifying sponge modules that can statistically distinguish between high-risk and low-risk patient groups.

Key Results

The application of miRspongeR to the BRCA dataset yielded several significant findings:

  • High-Confidence Interactions: The integrated method identified 87 robust miRNA sponge interactions.
  • Comparative Performance: An analysis of the overlap between methods showed that while miRHomology predicted the highest number of unique interactions, the sppc method had the highest validation rate against experimental data.
UpSet plot showing the intersection of miRNA sponge interactions identified by the seven different methods in the BRCA dataset.
  • Prognostic Value: Survival analysis identified modules with strong clinical relevance. For instance, modules found via the MCODE algorithm achieved p-values as low as $1.05 \times 10^{-6}$, highlighting their potential as cancer biomarkers.

Limitations and Future Directions

The primary challenge remains the high false-positive rate of initial miRNA-target prediction algorithms. As experimental "ground truth" databases expand, the accuracy of computational sponge prediction is expected to improve. Future iterations of miRspongeR will likely incorporate emerging identification methods to keep pace with the evolving field of ceRNA research.

Summary updated

Abstract

Background A microRNA (miRNA) sponge is an RNA molecule with multiple tandem miRNA response elements that can sequester miRNAs from their target mRNAs. Despite growing appreciation of the importance of miRNA sponges, our knowledge of their complex functions remains limited. Moreover, there is still a lack of miRNA sponge research tools that help researchers to quickly compare their proposed methods with other methods, apply existing methods to new datasets, or select appropriate methods for assisting in subsequent experimental design. Results To fill the gap, we present an R/Bioconductor package, miRspongeR, for simplifying the procedure of identifying and analyzing miRNA sponge interaction networks and modules. It provides seven popular methods and an integrative method to identify miRNA sponge interactions. Moreover, it supports the validation of miRNA sponge interactions and the identification of miRNA sponge modules, as well as functional enrichment and survival analysis of miRNA sponge modules. Conclusions This package enables researchers to quickly evaluate their new methods, apply existing methods to new datasets, and consequently speed up miRNA sponge research.
BibTeX
@article{zhang2019mirsponger,
  author = {Junpeng Zhang and Lin Liu and Taosheng Xu and Yong Xie and Chunwen Zhao and Jiuyong Li and Thuc Duy Le},
  title = {{miRspongeR: an R/Bioconductor package for the identification and analysis of miRNA sponge interaction networks and modules}},
  journal = {BMC Bioinformatics},
  year = {2019},
  doi = {10.1186/s12859-019-2861-y},
}
Figure