Bioinformatic Tools for the Analysis and Prediction of ncRNA Interactions

Abstract

Noncoding RNAs (ncRNAs) play prominent roles in the regulation of gene expression via their interactions with other biological molecules such as proteins and nucleic acids. Although much of our knowledge about how these ncRNAs operate in different biological processes has been obtained from experimental findings, computational biology can also clearly substantially boost this knowledge by suggesting possible novel interactions of these ncRNAs with other molecules. Computational predictions are thus used as an alternative source of new insights through a process of mutual enrichment because the information obtained through experiments continuously feeds through into computational methods. The results of these predictions in turn shed light on possible interactions that are subsequently validated experimentally. This review describes the latest advances in databases, bioinformatic tools, and new in silico strategies that allow the establishment or prediction of biological interactions of ncRNAs, particularly miRNAs and lncRNAs. The ncRNA species described in this work have a special emphasis on those found in humans, but information on ncRNA of other species is also included.

Keywords: bioinformatics; gene regulatory networks; genomics; interactome; lncRNA; ncRNA; transcriptome.

Figure 1

Classification of noncoding RNAs. The scheme presents the classification of noncoding RNAs according to their function, size, and location/orientation in the genome.

Figure 2

ncRNAs can regulate gene expression by diverse mechanisms. ncRNAs participate in the formation of nuclear bodies (1), gene transcription (2–3), modulate splicing events (4), regulate mRNA by degradation or stabilization (5), act as miRNA sponges (6), and ncRNAs can also be involved in the control of transcription (7) and cell signaling (8).

Figure 3

Interactions between noncoding RNAs and other molecules. (A) ncRNAs with other ncRNAs. Upper: miRNAs competing with lncRNA. Lower: CircRNA competing with miRNAs. (B) ncRNAs with mRNA. Upper: siRNAs silencing mRNA. Lower: Alternative splicing of mRNA due to an lncRNA. (C) ncRNAs with proteins. Upper: An lncRNA developing scaffold function and miRNAs activating Toll-like receptors. Lower: A circRNA serving as a sponge or Foxo3 protein. (D) ncRNAs with DNA. Upper: An lncRNA targeting the activator of a gene. Lower: An lncRNA altering the structure of DNA.

Figure 4

A computational approach for discovering or predicting RNA interactions among different biomolecules. The first strategy is to search in web tools or RNA databases such as miRTarBase. Another way to discover RNA interactions is to use resources based on deep learning and other machine learning-based tools, such as DeepTarget and deepMirGene. Finally, different mathematical and network theory methods can be used to research RNA interactions.