microRNA evolution in a human transcription factor and microRNA regulatory network
© Qiu et al; licensee BioMed Central Ltd. 2010
Received: 14 January 2010
Accepted: 29 June 2010
Published: 29 June 2010
microRNAs (miRNAs) are important cellular components. The understanding of their evolution is of critical importance for the understanding of their function. Although some specific evolutionary rules of miRNAs have been revealed, the rules of miRNA evolution in cellular networks remain largely unexplored. According to knowledge from protein-coding genes, the investigations of gene evolution in the context of biological networks often generate valuable observations that cannot be obtained by traditional approaches.
Here, we conducted the first systems-level analysis of miRNA evolution in a human transcription factor (TF)-miRNA regulatory network that describes the regulatory relations among TFs, miRNAs, and target genes. We found that the architectural structure of the network provides constraints and functional innovations for miRNA evolution and that miRNAs showed different and even opposite evolutionary patterns from TFs and other protein-coding genes. For example, miRNAs preferentially coevolved with their activators but not with their inhibitors. During transcription, rapidly evolving TFs frequently activated but rarely repressed miRNAs. In addition, conserved miRNAs tended to regulate rapidly evolving targets, and upstream miRNAs evolved more rapidly than downstream miRNAs.
In this study, we performed the first systems level analysis of miRNA evolution. The findings suggest that miRNAs have a unique evolution process and thus may have unique functions and roles in various biological processes and diseases. Additionally, the network presented here is the first TF-miRNA regulatory network, which will be a valuable platform of systems biology.
microRNAs (miRNAs) are a class of endogenous and small non-coding regulatory RNAs, which regulate genes at the post-transcriptional level . In the past few years, studies of miRNAs have ranged from their biological functions to their evolution. Understanding the evolution of miRNAs is very important to the study of their function, genomic organization, human disease, and medicine [2, 3]. Studies of miRNA evolution have focused on the molecular level. For example, the majority of miRNAs are conserved during evolution [1, 4, 5]. The structure of miRNA precursor stem loops exhibits significantly increased mutational robustness in comparison with random RNA sequences with the same stem-loop structure. It was revealed that the genetic robustness observed in miRNA sequences is the byproduct of selection for environmental robustness . Vazquez et al. found that recently evolved miRNAs consistently give rise to long-miRNAs, while ancient miRNAs give rise predominantly to canonical miRNAs in Arabidopsis . An Alu-mediated rapid expansion of miRNA genes in primate-specific miRNAs  and a rapid evolution of an X-linked miRNA cluster in primates were observed . Besides the fact that miRNAs are evolutionary conserved, it was observed that some miRNA genes are evolutionarily young . In addition, transposable elements (TE)-derived human miRNAs are less conserved, on average, than non-TE-derived miRNA s. The sequence diversification of duplicated miRNA genes to be accompanied by a change in spatial and temporal expression patterns . Host-virus coevolution may affect miRNA regulatory function . We previously found that miRNAs tend to buffer gene expression variation between closely related species  and human-specific miRNAs tend to evolve rapidly and found indications that some human miRNAs seem to be under recent positive selection . Lowly expressed human microRNA genes evolve rapidly . Recently, de Wit et al. revealed a novel mode of miRNA evolution, hairpin shifting . The above cited studies have uncovered some important evolutionary insights, but have, however, not considered the regulatory context of miRNAs. That is, as the primary two classes of gene regulators, miRNAs and transcription factors (TFs) regulate each other and co-regulate other genes. Therefore, considering the regulatory network of miRNAs in such evolutionary studies is increasingly important for an integrated understanding of the subject.
Cells typically change physiologically in response to signals received from their changing internal and external environments . To do this they must activate or repress the production of various gene products and tune these products to the proper level for different situations. Hence, the deregulation of genes may result in phenotypic variations that can contribute to diseases. For example, removing just one molecule of the transcription factor protein, c-myb, from the cellular milieu can result in developmental defects in the B cell lineage . The current research perspective is that the level of gene expression is primarily regulated by TFs at the transcriptional level and by miRNAs at the post-transcriptional level. Moreover, TFs and miRNAs can also regulate each other, and therefore they, together with their target genes, form a complex TF-miRNA regulatory network. Recent research has investigated the regulatory rules between miRNAs and genes, and found, for example, that miRNAs preferentially regulate genes that have a high transcriptional regulation complexity  and that preferentially target downstream genes in cellular signaling flows . These results support the concept that miRNA, TFs and their target genes form a complex network that enables the cell to conduct a wide range of biological functions. In light of this, studying miRNA evolution within the framework of cellular networks is essential.
At the molecular level, the topology of cellular networks places constraints on protein evolution and introduces functional innovations that open the door for protein evolution. The evolution of the protein-coding genes has been extensively studied in gene transcription [18, 22, 23], protein interaction , cell signaling , and metabolic  networks. These studies have led to several major conclusions: (1) Hub proteins, that is, proteins that have many interacting links, tend to be more conserved [18, 24, 25]. (2) Proteins in the network periphery tend to evolve more rapidly whereas those in the network center are more conserved . (3) Network proteins appear to have coevolved with their neighbors in a signaling network ; whereas transcription factors tended to evolve independently of their targets in prokaryotic transcriptional regulatory networks ; and so on. However, whether and how the architectural structure of cellular networks places constraints on and provides functional innovations for miRNA evolution is unknown. Did TFs coevolve with their target miRNAs and their target protein-coding genes? How does the evolution of miRNA-target pairs occur? Are the evolutionary rates of upstream regulators and downstream regulators different? Each of these questions is of critical importance not only for understanding evolution itself but also for related areas, such as the prediction of miRNA and TF targets.
To address these questions, we compared the evolution of miRNA and protein-coding genes in a manually curated TF-miRNA regulatory network. We used experimentally determined regulatory relations among TFs, miRNAs, and their targets to construct a human TF-miRNA regulatory network, which contains 2,273 nodes, of which 425 and 150 are TF and miRNA nodes, respectively. The network contains 4298 regulatory relations, including 2655 TF-gene regulatory relations, 210 TF-miRNA regulatory relations, and 1433 miRNA-target regulatory relations. We then performed a systems-level analysis to compare the evolutionary patterns of miRNAs and protein-coding genes in the network.
Results and Discussion
miRNAs preferentially coevolve with their activators but not with their inhibitors
We next investigated whether the signal type in the network contributes differently to the coevolution of TFs and protein-coding genes. Because the types of signal data of regulatory relationship between TFs and protein-coding genes are not available, we used the correlation coefficients of expression profiles of TFs and their target genes as an estimate of the signal type. A positive correlation may suggest an activating regulation; whereas a negative correlation may suggest a repressing regulation. As a result, the evolutionary rates of TF-gene pairs that have positive correlated expression profiles show positive correlation (R = 0.09, P = 0.017, Spearman's correlation), suggesting that TF-activating-gene pairs tend to show coevolution. As a comparison, TF-gene pairs that have negative correlated expression profiles do not show correlated evolutionary rates (R = -0.06, P = 0.249, Spearman's correlation).
Rapidly evolving TFs frequently activate but rarely repress miRNAs
We next asked whether there is a signal preference when TFs regulate miRNAs. We first classified TF-miRNA regulatory signals into two groups according to the evolutionary rates of their TFs and counted the numbers of activation signals and repression signals in these two groups. We found that 85.4% (76/89) of the signals in the high evolutionary rate group are activation signals, whereas this percentage decreases to 71.9% (82/114) in the low evolutionary rate group (P = 0.016, Fisher's exact test), suggesting that the signals are significantly unevenly distributed in these two groups. These findings indicate that rapidly evolving TFs are preferentially involved with transcriptional activation signals in TF-miRNA regulations.
We wondered whether a similar trend exists for the transcription regulation of protein-coding genes. Shinar et al. reported that protein-coding genes that are frequently needed in the natural environment tend to be activated but rarely needed genes tend to be repressed . In addition, broadly expressed protein-coding genes are more conserved than those with a narrow expression profile . Taken together, these findings indicate that conserved protein-coding genes tend to be activated by TFs and rapidly evolving protein-coding genes tend to be repressed. Because the signal type data of the protein-coding genes is not available, we were unable to test this trend in the human TF-miRNA regulatory network. When we used the estimated TF-gene regulatory signal type data from expression profile to perform this analysis, we did not obtained significant result (data not shown). However, it will be interesting to confirm this hypothesis when regulatory signal type data of human TF-gene regulations becomes available in the future.
Rapidly evolving and slowly evolving miRNAs tend to regulate slowly evolving and rapidly evolving protein-coding genes, respectively
For miRNA-gene regulatory pairs, we found that conserved miRNAs seem to regulate rapidly evolving protein-coding genes, whereas rapidly evolving miRNAs seem to regulate conserved protein-coding genes. We classified miRNA-target pairs into two equal groups according to the evolutionary rate of the miRNAs. We found that targets in the low rate group had a higher evolutionary rate than those in the high rate group (median dN: 0.052 vs. 0.046, P = 0.02, Wilcoxon test). To understand this phenomenon, we took into account the expression of miRNAs and their target genes. More conserved protein-coding genes tend to have a higher expression level and a larger breadth of expression . Recently Liang and Li reported a similar rule for miRNAs . Because miRNAs negatively regulate their target genes, broadly expressed (conserved) genes cannot be regulated by broadly expressed (conserved) miRNAs. If broadly expressed genes (that is, genes that are expressed in many tissues) are repressed by broadly expressed miRNAs, the broadly expressed protein-coding genes could be repressed. Thus, these broadly expressed genes could not continue to be broadly expressed. Therefore, broadly expressed genes tend to be regulated by tissue-specific miRNAs, which tend to evolve more rapidly.
In contrast, the coevolution between TF-protein-coding gene regulatory pairs is not significant (data not shown), which is consistent with the result of Madan Babu et al. .
Upstream miRNAs evolve more rapidly than downstream miRNAs, whereas upstream TFs are more conserved than downstream TFs
This result indicates that the two classes of gene regulators, TFs and miRNAs, show opposite patterns in evolutionary rates in the upstream and downstream of the regulatory cascade. The upstream TFs would seem to play more critical roles in the gene regulatory network than do downstream TFs because they not only regulate genes but also regulate other regulators (TFs and miRNAs). They tend to be more essential and therefore are more conserved than the downstream TFs because essential genes tend to more conserved, as Pal indicated .
On the other hand, miRNAs regulate genes by repressing them. The repressing function may have a systems-level function of buffering gene expression noise. For example, we previously showed that miRNAs buffer gene expression noise between species and thus buffer the evolution of the species . Wu and colleagues confirmed this function . In biological systems such as cell signaling it is desirable to filtering out noisy signals in the upstream region where genes are responding to a broad range of extracellular stimuli. An integrative analysis of the regulation of a human signaling network by miRNAs suggests that miRNAs could filter noisy signals in the upstream region of the signaling network . Furthermore, Legewie et al. suggested that negative feedbacks may serve as major regulatory loops in the upstream region of the signaling networks . Considering all of these aspects of miRNAs as native regulators, it is reasonable that upstream miRNAs evolve more rapidly than downstream miRNAs. Rapidly evolving miRNAs in the upstream of a regulatory cascade could allow adapt the cell to environmental changes and tone down the signaling process, as suggested by experimental studies in which signaling persisted if transcriptional feedback by proteins was blocked by protein biosynthesis inhibitors . More importantly, rapidly acting post-translational feedbacks may frequently be important for initial signal processing and specificity [29, 35]. Therefore, rapidly evolving miRNAs in the upstream of regulatory cascades allow the system to adapt in ways that allow for filtering out noisy signals and controlling the processing and specificity of the original signal.
Results of sensitivity analysis
Coevolution of TF-miRNA
R = 0.19, P = 0.003, Spearman's correlation
Coevolution of TF-activating-miRNA
R = 0.20, P = 0.01, Spearman's correlation
Coevolution of TF-repressing-miRNA
R = -0.04, P = 0.77, Spearman's correlation
Coevolution of TF-activating-gene
R = 0.08, P = 0.05, Spearman's correlation
Coevolution of TF-repressing-gene
R = -0.06, P = 0.27, Spearman's correlation
Evolutionary rates of targets regulated by rapidly evolving and slowly evolving miRNAs, respectively
Median dN 0.047 vs. 0.051, P = 0.05, Wilcoxon test
In summary, we have uncovered that the architectural structure of the TF-miRNA regulatory network provides constraints and functional innovations for miRNA evolution. Moreover, miRNAs have unique and even opposite evolutionary rules comparing with TFs and other protein-coding genes, suggesting that miRNAs may have unique functions and roles in various biological processes and diseases.
The most interesting discovery from this study is that the evolutionary patterns of miRNAs are different from those of the TFs and other protein-coding genes at the network level. For example, in the regulatory cascades, upstream miRNAs evolve more rapidly than downstream miRNAs, whereas upstream TFs are more conserved than downstream TFs. Most of the miRNAs are negative regulators. Taking both of these facts into consideration, we concluded that rapidly-evolving miRNAs in the upstream of the regulatory cascades allow the system to adapt in ways that allow the cell to filter out noisy signals and control the processing and specificity of the original signal. This conclusion supports the concept that miRNAs have buffers on their expression at the systems-level . Similarly, rapidly evolving and slowly evolving miRNAs tend to regulate slowly evolving and rapidly evolving protein-coding genes, respectively In addition, rapidly evolving TFs tend to activate miRNAs but tend to repress protein-coding genes. On the other hand, TFs and miRNAs have not evolved independently. For example, miRNAs preferentially coevolve with their activators (TFs), and rapidly evolving TFs preferentially activate miRNAs.
In addition, our findings will also be valuable for other fields, such as miRNA target prediction. We revealed that conserved miRNAs tend to avoid regulating conserved targets. This observation is helpful in designing better principles for the prediction of miRNA targets. Finally, the TF-miRNA regulatory network we presented in this study represents the first TF-miRNA regulatory network and will be a valuable platform of systems biology.
In this study, we performed an analysis of miRNA evolution in a human TF-miRNA regulatory network, which integrated the experimentally supported regulatory relations of TF-miRNA, TF-target, and miRNA-target. This network represents the first large-scale human TF-miRNA regulatory network. As a result, some principles and patterns of miRNA evolution in the human TF-miRNA regulatory network have been uncovered. These results are helpful for not only the understanding of miRNA origin, evolution, and function but also the development of novel methods for miRNA bioinformatics, for example the prediction of miRNA targets.
Construction and analysis of the human TF-miRNA regulatory network
We constructed a human TF-miRNA regulatory network based on experimentally supported regulatory relations between TFs and genes, between TFs and miRNAs, and between miRNAs and targets. We obtained the experimentally supported human TF-gene regulatory relations from TransFac (TRANSFAC Professional version, January 2009, http://www.gene-regulation.com) and the experimentally supported human miRNA-target regulatory relations from TarBase . In order to obtain experimentally supported TF-miRNA regulatory relations, we manually curated ~5000 papers published before April 2009 and obtained experimentally supported TF-miRNA regulatory relations http://cmbi.bjmu.edu.cn/transmir. We next constructed a human TF-miRNA regulatory network using the above three types of regulatory relations among TFs, miRNAs, and target genes (Additional file 1). Of the TF-miRNA links, activating links is greatly more than, repressing links. This result indicates that, for the transcription of miRNAs, activation interactions are more common than repression interactions, a finding which is consistent with the observation for the transcription of protein-coding genes . Normally miRNAs act as negative gene regulators and all of the miRNA-target links identified in this study are indeed negative. The network nodes have a skewed degree distribution. Specifically, most of the TFs regulate just a few miRNAs, but some TFs regulate many miRNAs. For example, the TF that regulats the largest number of miRNAs in this network, MYC, regulates 26 miRNAs.
We implemented a Java program to identify the components of this network (Additional file 2). The resulting network contains 29 network components, of which the largest network component contains 97% of the network nodes. This finding suggests that TFs, miRNAs, and target genes intereact in a single, closely interconnected TF-miRNA-target regulatory network. We identified upstream and downstream TFs/miRNAs using shortest paths, which is obtained by Dijkstra's algorithm.
Evolutionary rate data of human genes and miRNAs
In this study, we used dN as a measure of the evolutionary rate of protein-coding genes  and used the miRNA sequence divergence as the measure of the evolutionary rate of miRNAs . We downloaded the human-mouse protein dN data (Additional file 3) from the H-InvDB database http://jbirc.jbic.or.jp/hinv/dataset/download.cgi. In order to calculate the evolutionary rates for human miRNAs, we first downloaded the pairwise alignment data for humans (hg18) and Rhesus monkeys (rheMac2) from UCSC . We next obtained the genome coordinates data for known human miRNAs from miRBase. We then calculated the sequence divergence for human miRNAs (Additional file 4) using the method presented by Liang et al .
We performed a Spearman's correlation test, Wilcoxon test, and Fisher's exact test using R, a statistical computing language http://www.r-project.org/.
We thank the reviewers for their valuable comments and suggestions. We thank Drs. Edmund F. and Rhoda E. Perozzi for English editing assistance. This work was supported by the Natural Science Foundation of China (Grant No. 30900829) and was supported by Doctoral Fund of Ministry of Education of China (Grant No. 20090001120040).
- Berezikov E, Cuppen E, Plasterk RH: Approaches to microRNA discovery. Nat Genet. 2006, 38 (Suppl): S2-7. 10.1038/ng1794View ArticlePubMedGoogle Scholar
- Lu M, Zhang Q, Deng M, Miao J, Guo Y, Gao W, Cui Q: An analysis of human microRNA and disease associations. PLoS ONE. 2008, 3 (10): e3420-doi:10.1371/journal.pone.0003420,PubMed CentralView ArticlePubMedGoogle Scholar
- Zhang Q, Lu M, Cui Q: SNP analysis reveals an evolutionary acceleration of the human-specific microRNAs. Nature Precedings. 2008, Available from http://hdlhandlenet/10101/npre200821271Google Scholar
- Chen K, Rajewsky N: Natural selection on human microRNA binding sites inferred from SNP data. Nat Genet. 2006, 38 (12): 1452-1456. 10.1038/ng1910View ArticlePubMedGoogle Scholar
- Saunders MA, Liang H, Li WH: Human polymorphism at microRNAs and microRNA target sites. Proc Natl Acad Sci USA. 2007, 104 (9): 3300-3305. 10.1073/pnas.0611347104PubMed CentralView ArticlePubMedGoogle Scholar
- Borenstein E, Ruppin E: Direct evolution of genetic robustness in microRNA. Proc Natl Acad Sci USA. 2006, 103 (17): 6593-6598. 10.1073/pnas.0510600103PubMed CentralView ArticlePubMedGoogle Scholar
- Szollosi GJ, Derenyi I: Congruent evolution of genetic and environmental robustness in micro-RNA. Mol Biol Evol. 2009, 26 (4): 867-874. 10.1093/molbev/msp008View ArticlePubMedGoogle Scholar
- Vazquez F, Blevins T, Ailhas J, Boller T, Meins F, Jr : Evolution of Arabidopsis MIR genes generates novel microRNA classes. Nucleic Acids Res. 2008, 36 (20): 6429-6438. 10.1093/nar/gkn670PubMed CentralView ArticlePubMedGoogle Scholar
- Zhang R, Wang YQ, Su B: Molecular evolution of a primate-specific microRNA family. Mol Biol Evol. 2008, 25 (7): 1493-1502. 10.1093/molbev/msn094View ArticlePubMedGoogle Scholar
- Zhang R, Peng Y, Wang W, Su B: Rapid evolution of an X-linked microRNA cluster in primates. Genome Res. 2007, 17 (5): 612-617. 10.1101/gr.6146507PubMed CentralView ArticlePubMedGoogle Scholar
- Lu J, Shen Y, Wu Q, Kumar S, He B, Shi S, Carthew RW, Wang SM, Wu CI: The birth and death of microRNA genes in Drosophila. Nat Genet. 2008, 40 (3): 351-355. 10.1038/ng.73View ArticlePubMedGoogle Scholar
- Piriyapongsa J, Marino-Ramirez L, Jordan IK: Origin and evolution of human microRNAs from transposable elements. Genetics. 2007, 176 (2): 1323-1337. 10.1534/genetics.107.072553PubMed CentralView ArticlePubMedGoogle Scholar
- Maher C, Stein L, Ware D: Evolution of Arabidopsis microRNA families through duplication events. Genome Res. 2006, 16 (4): 510-519. 10.1101/gr.4680506PubMed CentralView ArticlePubMedGoogle Scholar
- Berezikov E, Plasterk RH: Camels and zebrafish, viruses and cancer: a microRNA update. Hum Mol Genet. 2005, 14: R183-190. Spec No. 2, 10.1093/hmg/ddi271View ArticlePubMedGoogle Scholar
- Cui Q, Yu Z, Purisima EO, Wang E: MicroRNA regulation and interspecific variation of gene expression. Trends Genet. 2007, 23 (8): 372-375. 10.1016/j.tig.2007.04.003View ArticlePubMedGoogle Scholar
- Liang H, Li WH: Lowly expressed human microRNA genes evolve rapidly. Mol Biol Evol. 2009, 26 (6): 1195-1198. 10.1093/molbev/msp053PubMed CentralView ArticlePubMedGoogle Scholar
- de Wit E, Linsen SE, Cuppen E, Berezikov E: Repertoire and evolution of miRNA genes in four divergent nematode species. Genome Res. 2009,Google Scholar
- Perez JC, Groisman EA: Evolution of transcriptional regulatory circuits in bacteria. Cell. 2009, 138 (2): 233-244. 10.1016/j.cell.2009.07.002PubMed CentralView ArticlePubMedGoogle Scholar
- Xiao C, Calado DP, Galler G, Thai TH, Patterson HC, Wang J, Rajewsky N, Bender TP, Rajewsky K: MiR-150 controls B cell differentiation by targeting the transcription factor c-Myb. Cell. 2007, 131 (1): 146-159. 10.1016/j.cell.2007.07.021View ArticlePubMedGoogle Scholar
- Cui Q, Yu Z, Pan Y, Purisima EO, Wang E: MicroRNAs preferentially target the genes with high transcriptional regulation complexity. Biochem Biophys Res Commun. 2007, 352 (3): 733-738. 10.1016/j.bbrc.2006.11.080View ArticlePubMedGoogle Scholar
- Cui Q, Yu Z, Purisima EO, Wang E: Principles of microRNA regulation of a human cellular signaling network. Mol Syst Biol. 2006, 2: 46- 10.1038/msb4100089PubMed CentralView ArticlePubMedGoogle Scholar
- Jothi R, Balaji S, Wuster A, Grochow JA, Gsponer J, Przytycka TM, Aravind L, Babu MM: Genomic analysis reveals a tight link between transcription factor dynamics and regulatory network architecture. Mol Syst Biol. 2009, 5: 294- 10.1038/msb.2009.52PubMed CentralView ArticlePubMedGoogle Scholar
- Madan Babu M, Teichmann SA, Aravind L: Evolutionary dynamics of prokaryotic transcriptional regulatory networks. J Mol Biol. 2006, 358 (2): 614-633. 10.1016/j.jmb.2006.02.019View ArticlePubMedGoogle Scholar
- Fraser HB, Hirsh AE, Steinmetz LM, Scharfe C, Feldman MW: Evolutionary rate in the protein interaction network. Science. 2002, 296 (5568): 750-752. 10.1126/science.1068696View ArticlePubMedGoogle Scholar
- Cui Q, Purisima EO, Wang E: Protein evolution on a human signaling network. BMC Syst Biol. 2009, 3: 21- 10.1186/1752-0509-3-21PubMed CentralView ArticlePubMedGoogle Scholar
- Alves R, Chaleil RA, Sternberg MJ: Evolution of enzymes in metabolism: a network perspective. J Mol Biol. 2002, 320 (4): 751-770. 10.1016/S0022-2836(02)00546-6View ArticlePubMedGoogle Scholar
- Karolchik D, Hinrichs AS, Furey TS, Roskin KM, Sugnet CW, Haussler D, Kent WJ: The UCSC Table Browser data retrieval tool. Nucleic Acids Res. 2004, 32 (Database issue): D493-496. 10.1093/nar/gkh103PubMed CentralView ArticlePubMedGoogle Scholar
- Cui Q, Ma Y, Jaramillo M, Bari H, Awan A, Yang S, Zhang S, Liu L, Lu M, O'Connor-McCourt M, et al.: A map of human cancer signaling. Mol Syst Biol. 2007, 3: 152- 10.1038/msb4100200PubMed CentralView ArticlePubMedGoogle Scholar
- Fu C, Li J, Wang E: Signaling network analysis of ubiquitin-mediated proteins suggests correlations between the 26S proteasome and tumor progression. Mol Biosyst. 2009, 5 (12): 1809-1816. 10.1039/b905382dView ArticlePubMedGoogle Scholar
- Shinar G, Dekel E, Tlusty T, Alon U: Rules for biological regulation based on error minimization. Proc Natl Acad Sci USA. 2006, 103 (11): 3999-4004. 10.1073/pnas.0506610103PubMed CentralView ArticlePubMedGoogle Scholar
- Pal C, Papp B, Lercher MJ: An integrated view of protein evolution. Nat Rev Genet. 2006, 7 (5): 337-348. 10.1038/nrg1838View ArticlePubMedGoogle Scholar
- Wu CI, Shen Y, Tang T: Evolution under canalization and the dual roles of microRNAs: a hypothesis. Genome Res. 2009, 19 (5): 734-743. 10.1101/gr.084640.108PubMed CentralView ArticlePubMedGoogle Scholar
- Legewie S, Herzel H, Westerhoff HV, Bluthgen N: Recurrent design patterns in the feedback regulation of the mammalian signalling network. Mol Syst Biol. 2008, 4: 190- 10.1038/msb.2008.29PubMed CentralView ArticlePubMedGoogle Scholar
- Amit I, Wides R, Yarden Y: Evolvable signaling networks of receptor tyrosine kinases: relevance of robustness to malignancy and to cancer therapy. Mol Syst Biol. 2007, 3: 151- 10.1038/msb4100195PubMed CentralView ArticlePubMedGoogle Scholar
- Altan-Bonnet G, Germain RN: Modeling T cell antigen discrimination based on feedback control of digital ERK responses. PLoS Biol. 2005, 3 (11): e356- 10.1371/journal.pbio.0030356PubMed CentralView ArticlePubMedGoogle Scholar
- Papadopoulos GL, Reczko M, Simossis VA, Sethupathy P, Hatzigeorgiou AG: The database of experimentally supported targets: a functional update of TarBase. Nucleic Acids Res. 2009, 37 (Database issue): D155-158. 10.1093/nar/gkn809PubMed CentralView ArticlePubMedGoogle Scholar
- Wang J, Lu M, Qiu C, Cui Q: TransmiR: a transcription factor-microRNA regulation database. Nucleic Acids Res. 2010, 38 (Database issue): D119-122. 10.1093/nar/gkp803PubMed CentralView ArticlePubMedGoogle Scholar
- Shen-Orr SS, Milo R, Mangan S, Alon U: Network motifs in the transcriptional regulation network of Escherichia coli. Nat Genet. 2002, 31 (1): 64-68. 10.1038/ng881View ArticlePubMedGoogle Scholar
- Zhou T, Drummond DA, Wilke CO: Contact density affects protein evolutionary rate from bacteria to animals. J Mol Evol. 2008, 66 (4): 395-404. 10.1007/s00239-008-9094-4View ArticlePubMedGoogle Scholar