[Paper Review] Characteristics Of Intronic And Intergenic Human Mirnas And Features Of Their Interaction With Mrna

Assel Issabekova, Olga Berillo+4 more|arXiv (Cornell University)|Nov 24, 2011

MicroRNA in disease regulation44 references6 citations

TL;DR

This study computationally predicts and characterizes miRNA binding sites across 51 human oncogenes, revealing that 5'UTR and CDS regions harbor significantly higher miRNA site density than 3'UTR. Using RNAHybrid and custom scripts, it identifies three interaction types—5'-dominant canonical, 3'-compensatory, and complementary—demonstrating that intronic miRNAs do not target host gene mRNAs, while intergenic and intronic miRNAs show distinct regulatory linkages in gastrointestinal cancer pathways.

ABSTRACT

Regulatory relationships of 686 intronic miRNA and 784 intergenic miRNAs with mRNAs of 51 intronic miRNA coding genes were established. Interaction features of studied miRNAs with 5'UTR, CDS and 3'UTR of mRNA of each gene were revealed. Functional regions of mRNA were shown to be significantly heterogenous according to the number of binding sites of miRNA and to the location density of these sites.

Motivation & Objective

• To identify and characterize miRNA binding sites in 5'UTR, CDS, and 3'UTR of oncogenes involved in gastrointestinal and breast cancer.
• To determine the distribution and density of binding sites for intergenic (ig-miRNA), intronic (in-miRNA), and exonic (ex-miRNA) miRNAs.
• To investigate the structural and energetic features of miRNA:mRNA interactions, including seed-dependent and compensatory mechanisms.
• To explore regulatory linkages between genes encoding miRNAs and their target mRNAs, particularly in cancer-related pathways.
• To validate the functional relevance of miRNA binding sites using thermodynamic energy (ΔG/ΔGm) and statistical significance (p < 0.0005).

Proposed method

• Used miRBase and GenBank to retrieve miRNA and mRNA sequences (Homo sapiens, GRCh37.2).
• Applied miRNAFinder 2.2 to classify miRNAs as intergenic, intronic, or exonic.
• Employed RNAHybrid 2.1 to predict miRNA:mRNA binding sites, including position, binding energy (ΔG), and interaction type.
• Developed E-RNAhybrid script to compute ΔG/ΔGm ratio (percent of maximum binding energy), p-values, and equalizing coefficients based on miRNA length.
• Calculated site density as sites per 1000 nucleotides (s/l) in 5'UTR, CDS, and 3'UTR regions.
• Set significance threshold at p < 0.0005 and used GC-content and miRNA length (21 nt as baseline) to normalize predictions.

Experimental results

Research questions

• RQ1 Where are the highest densities of miRNA binding sites located across 5'UTR, CDS, and 3'UTR regions of oncogene mRNAs?
• RQ2 Do intronic miRNAs (in-miRNAs) regulate the expression of their host gene mRNAs?
• RQ3 What are the dominant types of miRNA:mRNA interaction mechanisms (e.g., 5'-dominant, 3'-compensatory, complementary) in human oncogenes?
• RQ4 How do intergenic and intronic miRNAs contribute to regulatory networks among cancer-related genes?
• RQ5 What is the role of miRNA length and GC-content in determining binding site specificity and reliability?

Key findings

• 5'UTR has 2.6 times higher miRNA site density than CDS and 2.8 times higher than 3'UTR, with 53.1% of sites in CDS and 15.8% in 5'UTR across 51 oncogenes.
• Intronic miRNAs (in-miRNAs) show no significant binding to mRNAs of their host genes, indicating no autoregulation.
• 30.8% of miRNA sites are in 5'UTR, 53.1% in CDS, and 31.1% in 3'UTR, challenging the assumption that miRNA regulation is predominantly 3'UTR-driven.
• The average number of miRNA binding sites per mRNA is 6, with in-miRNAs showing 19% more sites per mRNA than ig-miRNAs.
• Three distinct interaction types were identified: 5'-dominant canonical (seed-dependent), 3'-compensatory, and fully complementary, with contributions from all miRNA regions to binding energy (ΔG/ΔGm).
• miRNAs with 22 nt length and 50–55% GC-content are most prevalent, and the equalizing coefficient improves prediction accuracy for non-21 nt miRNAs.