Table 1.
Summary of model components/reactions.
Figure 1.
A: The model of the EGFR signaling pathway contains 4 sub-pathways: MAPK pathway, PI3K-AKT pathway, Ca2+ signaling pathway, PAK signaling pathway. B, C are parts of the model network visualized by PyBioS. B: The mRNA BAD-1 is produced by the transcription reaction (1) of gene BAD and also takes part in decay reaction (2) and translation-reaction (3), which produces the protein BAD-1 (cytoplasm). The protein BAD-1 takes part in further three reactions that are the decay-reaction (4), phosphorylation reaction (5) and dephosphorylation reaction (6). The phosphorylated protein P-AKT (plasma membrane) catalyzes the phosphorylation reaction, in which protein BAD-1 is phosphorylated into P-BAD-1. Afterwards, the protein P-BAD-1 is then dephosphorylated; C: shows a simplified miRNA biogenesis, target recognition and competitive anti-miRNA effect. (1) miRNA-gene transcription; (2) miRNA translocation (from nucleus into cytoplasm); (3) miRNA-binding-target reaction; (4) miRNA binds to the miRNA inhibitor.
Figure 2.
Comparison of model predictions with experimental results.
A: Experimental results of relative concentration changes of target mRNAs according to individual miRNA over-expression experiments from the Avraham's study [30] (Reprinted with permission from AAAS); B: in silico prediction result of relative concentration changes of target-mRNAs according to each miRNA over-expression in the EGFR model. Both heatmaps show very similar qualitative results (protein down-regulation), the only discrepancies are for miR-155 and miR-498. mRNAs with low concentration changes (log2-ratio<0.001) are ignored and shown in ‘white’.
Figure 3.
Modeling of mir-192 and mir-181c effects on the EGFR signaling pathway.
A: Increased expression of mir-192 gene corresponds with a reduced level of targets' protein expression; B: Concentrations of EGFR downstream activated proteins are inversely correlated with the mir-192 gene expression level; C: simulation result with fixed mir-181c gene expression level (1 nM), whereas all other 240 miRNA genes are not expressed. All 13 AKT-dependent proteins can only be activated after the concentration of EGFR protein is getting higher than 10 pM. This activation threshold is due to the presence of mir-181c. D: represents another simulation result without the presence of mir-181c, additionally all other 240 miRNAs were not expressed, as well. At this condition, the activation threshold of these proteins is at 0.001 pM. By comparing these two results (C, D), one can understand that mir-181c raises the activation threshold of the EGFR signaling pathway significantly.
Figure 4.
Modeling of anti-mir-489 and anti-mir-34a effects on the EGFR signaling pathway.
Heatmaps A and B are the results of anti-mir-489 simulations. C and D are the results of anti-mir-34a simulations. The first experiment within each heatmap (from bottom to top) is the ‘control’ state; the miRNA effect and anti-miRNA effect of the different experiments is always compared versus the ‘control’ state. A: Quantitative changes of mir-489 direct target-proteins due to different amounts of anti-mir-489 per experiment. The inhibition effects of mir-489 is inversely correlated with the anti-mir-489 concentration. B: Quantitative changes of the EGFR downstream activated proteins according to different amounts of anti-mir-489 per experiment. The concentrations of many downstream activated proteins of the EGFR signaling pathway correlate with the concentration of anti-mir-489. C: Quantitative changes of mir-34a direct target-proteins due to different amounts of anti-mir-34a per experiment. The inhibition effects of mir-34a inversely correlate with the anti-mir-34a concentration. D: Quantitative changes of the EGFR downstream activated proteins according to different amounts of anti-mir-34a per experiment. The concentrations of many downstream activated proteins correlate with the anti-mir-34a concentration.
Figure 5.
Modeling the individual effect of 100 anti-miRNAs.
A and B: Each anti-miRNA is activated when the corresponding miRNA is over-expressed individually (an anti-mir-combination effect is omitted). Each row in this heatmap represents the predicted anti-miRNA effect and the more columns in the same row appears red or orange, the stronger is the predicted effect that this particular miRNA inhibitor can exert on the EGFR signaling pathway. In this manner, we can examine the impact of each anti-miRNA on this signaling pathway.
Table 2.
Top 15 anti-miRNAs.
Figure 6.
Histogram of miRNA/target relationships.
A: The top ranked 15 miRNAs of Table 2 correlated with the amount of their corresponding targets in the EGFR signaling pathway (Table 1). B: The 15 miRNAs correlated with amount of their corresponding targets in EGFR signaling pathway. The anti-miRNAs of these miRNAs have much less effect on this pathway (Fig. 5A and B).
Figure 7.
Transition fire of Petri Nets.
An exemplary phosphorylation reaction of the Petri Nets. BAD-1 (p1), ATP (p2), P-BAD-1 (p3), ADP (p4) and P-AKT (p5) are places and phosphorylation reaction (t1) is a transition. A: the concentrations of model components at time point 0 (before transition fire); B: the concentrations of components in model at time point 1 (after transition fire).
Table 3.
Transition t1 at time point 0 (before firing).
Table 4.
Transition t1 at time point 1 (after firing).
Table 5.
The summary of kinetic parameters applied in the Petri nets simulation.