Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Comparative Study
, 8, 396

Inter- And Intra-Combinatorial Regulation by Transcription Factors and microRNAs

Affiliations
Comparative Study

Inter- And Intra-Combinatorial Regulation by Transcription Factors and microRNAs

Yiming Zhou et al. BMC Genomics.

Abstract

Background: MicroRNAs (miRNAs) are a novel class of non-coding small RNAs. In mammalian cells, miRNAs repress the translation of messenger RNAs (mRNAs) or degrade mRNAs. miRNAs play important roles in development and differentiation, and they are also implicated in aging, and oncogenesis. Predictions of targets of miRNAs suggest that they may regulate more than one-third of all genes. The overall functions of mammalian miRNAs remain unclear. Combinatorial regulation by transcription factors alone or miRNAs alone offers a wide range of regulatory programs. However, joining transcriptional and post-transcriptional regulatory mechanisms enables higher complexity regulatory programs that in turn could give cells evolutionary advantages. Investigating coordinated regulation of genes by miRNAs and transcription factors (TFs) from a statistical standpoint is a first step that may elucidate some of their roles in various biological processes.

Results: Here, we studied the nature and scope of coordination among regulators from the transcriptional and miRNA regulatory layers in the human genome. Our findings are based on genome wide statistical assessment of regulatory associations ("interactions") among the sets of predicted targets of miRNAs and sets of putative targets of transcription factors. We found that combinatorial regulation by transcription factor pairs and miRNA pairs is much more abundant than the combinatorial regulation by TF-miRNA pairs. In addition, many of the strongly interacting TF-miRNA pairs involve a subset of master TF regulators that co-regulate genes in coordination with almost any miRNA. Application of standard measures for evaluating the degree of interaction between pairs of regulators show that strongly interacting TF-miRNA, TF-TF or miRNA-miRNA pairs tend to include TFs or miRNAs that regulate very large numbers of genes. To correct for this potential bias we introduced an additional Bayesian measure that incorporates not only how significant an interaction is but also how strong it is. Putative pairs of regulators selected by this procedure are more likely to have biological coordination. Importantly, we found that the probability of a TF-miRNA pair forming feed forward loops with its common target genes (where the miRNA simultaneously suppresses the TF and many of its targets) is increased for strongly interacting TF-miRNA pairs.

Conclusion: Genes are more likely to be co-regulated by pairs of TFs or pairs of miRNAs than by pairs of TF-miRNA, perhaps due to higher probability of evolutionary duplication events of shorter DNA sequences. Nevertheless, many gene sets are reciprocally regulated by strongly interacting pairs of TF-miRNA, which suggests an effective mechanism to suppress functionally related proteins. Moreover, the particular type of feed forward loop (with two opposing modes where the TF activates its target genes or the miRNA simultaneously suppresses this TF and the TF-miRNA joint target genes) is more prevalent among strongly interacting TF-miRNA pairs. This may be attributed to a process that prevents waste of cellular resources or a mechanism to accelerate mRNA degradation.

Figures

Figure 1
Figure 1
TF and miRNA interaction heatmap. Each pixel on Figures 1a and 1b represents the association of a unique pair of regulators. Figure 1a measures this association by a Fisher's Exact Test p-value (dark pixels represent lower p-values or alternatively a higher value of -log10p). Figure 1b measures the association by the Bayesian probability Pr{logOR>0.6} (Here a dark pixel means a high probability). The TFs and miRNAs are ordered so that the number of targets of each regulator increases as one moves across the Figure from left to right on the horizontal axis, and also up the vertical axis. Both Figures 1a and 1b illustrate that while TF-TF and miRNA-miRNA associations are common, TF-miRNA interactions are less so. The TF-miRNA rectangles of Fig 1a demonstrate that the most significant associations (as found by Fisher's Exact Test) tend to involve TF-miRNA pairs with the TF having a large number of targets. In the corresponding areas of Figure 1b, we see a more uniform sprinkling of dark points, indicating that the Bayesian approach is less sensitive to sample size effects. The stripes on the TF-miRNA rectangles of both figures demonstrate that certain TFs are associated with almost all the miRNAs – while, surprisingly, many TFs with a similar number of targets seem to not be significantly associated with any miRNA.
Figure 2
Figure 2
Comparison of Fisher's exact test and Bayesian association score. Scatter plots of Fisher's Exact Test p-value as a function of Bayesian association score. The 2D distributions demonstrate how the relationship between Fisher's Exact Test and our Bayesian score depends on the logOR threshold we use. Each sub-plot represents a different threshold value ranging from 0 to 1 – as indicated by each subtitle. For a particular threshold value, a pixel on the plot represents the local density of miRNA-TF pairs having the corresponding p-value (from the y-axis) and Bayesian Probability (from the x-axis). Here darker shaded regions indicate higher densities. For a 0 threshold, the Bayesian Test and Fisher's Test agree exactly. As we increase the threshold, we see fewer and fewer TF-miRNA pairs that are highly associated as measured by both ranking criteria (pairs whose measures approach 1 at the x-axis and 0 at the y-axis). The higher the threshold, the more emphasis we are placing on the size of the TF-miRNA association (as measured by a log Odds Ratio) and the less emphasis on sample size. Note that a very high Bayesian probability implies that the associated p-value will be small, no matter what threshold we use.
Figure 3
Figure 3
Feed Forward Loop (FFL). A feed forward loop (FFL) is a regulatory motif in which regulator A regulates another regulator denoted by B, and both regulators A and B regulate a common target C.
Figure 4
Figure 4
Relationship between FFL and TF/miRNA association. Fraction of FFLs as a function of the statistical significance for TFs and miRNAs association. The histogram displays the fraction of FFLs that result in each bin, when grouping miRNA/TF/GO triplets according to their log p-value of joint-association. To generate this histogram, we used a slightly restricted set of biological-process GO terms, such that each group includes at least one gene that is a predicted target of a TF and a miRNA. The plot suggests that when a miRNA/TF/GO triplet is significantly associated, the corresponding miRNA and TF are more likely to form a feed forward loop.

Similar articles

See all similar articles

Cited by 52 PubMed Central articles

See all "Cited by" articles

References

    1. Ambros V, Horvitz HR. Heterochronic mutants of the nematode Caenorhabditis elegans. Science. 1984;226:409–416. doi: 10.1126/science.6494891. - DOI - PubMed
    1. Lee RC, Feinbaum RL, Ambros V. The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell. 1993;75:843–854. doi: 10.1016/0092-8674(93)90529-Y. - DOI - PubMed
    1. Wightman B, Ha I, Ruvkun G. Posttranscriptional regulation of the heterochronic gene lin-14 by lin-4 mediates temporal pattern formation in C. elegans. Cell. 1993;75:855–862. doi: 10.1016/0092-8674(93)90530-4. - DOI - PubMed
    1. Borchert GM, Lanier W, Davidson BL. RNA polymerase III transcribes human microRNAs. Nat Struct Mol Biol. 2006;13:1097–1101. doi: 10.1038/nsmb1167. - DOI - PubMed
    1. Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell. 2004;116:281–297. doi: 10.1016/S0092-8674(04)00045-5. - DOI - PubMed

Publication types

LinkOut - more resources

Feedback