MiRNA-Mediated Relationships Between Cis-SNP Genotypes and Transcript Intensities in Lymphocyte Cell Lines.

Funding Source

NIH, NSF, Louisian BOR award

Grant Number

NCRR P20RR016456, RCMI 5G12RR026260-03, R21LM010137S1 to DZ, ZK (LEQSF(2008-11)-RD-A-32) EPS-1006891,


Department of Physics and Computer Science - Dual Degree Engineering

Document Type


Publication Date



In metazoans, miRNAs regulate gene expression primarily through binding to target sites in the 3′ UTRs (untranslated regions) of messenger RNAs (mRNAs). Cis-acting variants within, or close to, a gene are crucial in explaining the variability of gene expression measures. Single nucleotide polymorphisms (SNPs) in the 3′ UTRs of genes can affect the base-pairing between miRNAs and mRNAs, and hence disrupt existing target sites (in the reference sequence) or create novel target sites, suggesting a possible mechanism for cis regulation of gene expression. Moreover, because the alleles of different SNPs within a DNA sequence of limited length tend to be in strong linkage disequilibrium (LD), we hypothesize the variants of miRNA target sites caused by SNPs potentially function as bridges linking the documented cis-SNP markers to the expression of the associated genes. A large-scale analysis was herein performed to test this hypothesis. By systematically integrating multiple latest information sources, we found 21 significant gene-level SNP-involved miRNA-mediated post-transcriptional regulation modules (SNP-MPRMs) in the form of SNP-miRNA-mRNA triplets in lymphocyte cell lines for the CEU and YRI populations. Among the cognate genes, six including ALG8, DGKE, GNA12, KLF11, LRPAP1, and MMAB are related to multiple genetic diseases such as depressive disorder and Type-II diabetes. Furthermore, we found that ~35% of the documented transcript intensity-related cis-SNPs (~950) in a recent publication are identical to, or in significant linkage disequilibrium (LD) (p<0.01) with, one or multiple SNPs located in miRNA target sites. Based on these associations (or identities), 69 significant exon-level SNP-MPRMs and 12 disease genes were further determined for two populations. These results provide concrete in silico evidence for the proposed hypothesis. The discovered modules warrant additional follow-up in independent laboratory studies.


DOI: 10.1371/journal.pone.0031429.

Funding: This work was supported by NIH grants (NCRR P20RR016456, RCMI 5G12RR026260-03, R21LM010137S1 to DZ), a Louisiana BOR award to ZK (LEQSF(2008-11)-RD-A-32) and an NSF grant EPS-1006891). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.