Differential evolutionary conservation of motif modes in the yeast protein interaction network
© Lee et al; licensee BioMed Central Ltd. 2006
Received: 10 January 2006
Accepted: 25 April 2006
Published: 25 April 2006
The importance of a network motif (a recurring interconnected pattern of special topology which is over-represented in a biological network) lies in its position in the hierarchy between the protein molecule and the module in a protein-protein interaction network. Until now, however, the methods available have greatly restricted the scope of research. While they have focused on the analysis in the resolution of a motif topology, they have not been able to distinguish particular motifs of the same topology in a protein-protein interaction network.
We have been able to assign the molecular function annotations of Gene Ontology to each protein in the protein-protein interactions of Saccharomyces cerevisiae. For various motif topologies, we have developed an algorithm, enabling us to unveil one million "motif modes", each of which features a unique topological combination of molecular functions. To our surprise, the conservation ratio, i.e., the extent of the evolutionary constraints upon the motif modes of the same motif topology, varies significantly, clearly indicative of distinct differences in the evolutionary constraints upon motifs of the same motif topology. Equally important, for all motif modes, we have found a power-law distribution of the motif counts on each motif mode. We postulate that motif modes may very well represent the evolutionary-conserved topological units of a protein interaction network.
For the first time, the motifs of a protein interaction network have been investigated beyond the scope of motif topology. The motif modes determined in this study have not only enabled us to differentiate among different evolutionary constraints on motifs of the same topology but have also opened up new avenues through which protein interaction networks can be analyzed.
In the post-genomic era, one major goal of functional genomics has been to identify and analyze molecular interactions in a cellular context to better understand the mechanisms according to which biological molecules interact and function. A protein-protein interaction (PPI) is a specific type of molecular interaction that plays the central role in relaying signals, in building molecular machines, in engaging in enzyme reactions and in decision-making vis-à-vis multiple biological processes. Advancements in PPI-detection technology have unquestionably led to the rapid accumulation of PPI data [1–5]. To accurately conceptualize a PPI by determining its precise description is to immediately allow for the utilization of the various tools currently accessible in network science . A PPI network, for instance, has been found to have a scale-free structure; i.e., the link count (interactions) of a protein (node) follows a broad-tailed distribution that is approximated as a power-law,P(k)~k -γ , where k is the link count and γ is the degree exponent [6, 7, 7–9]. A network diameter, defined as the average minimal path between two nodes, is reportedly small (~6–7) in a PPI network [9, 10].
The contribution made by research on network motifs, where a network motif is the specific topology of a combination of nodes that occurs repeatedly at different positions in that network, is evidenced in the results obtained from the initial search for cause of the topological properties. As the first step, one of the properties of a PPI network, the scale-free feature, was thoroughly examined [11, 12]. At issue was whether the selection force is the cause of one of the scale-free network features: a random mutation does not harm a network, as a whole, but can cause it to collapse, but only with a deliberate attack on the hubs (nodes that contain many immediate neighbors) . It was, therefore, reasonably inferred and later clearly observed that hubs are more likely to evolve at a slower rate than non-hub nodes [14, 15]. However, different conclusions have been drawn regarding the mutation rate of a hub [12, 16]. In addition to this, recent research into gene regulatory networks in yeast has uncovered both a power-law distribution for the number of regulated genes per regulating protein  and, at the same time, has yielded invaluable information concerning the presence of typical patterns of motifs. Feed forward motifs, single-input motifs and dense overlapping regulons [18–21] reportedly occur in the gene regulatory network (transcription factor vs. the regulated gene) of Saccharomyces cerevisiae and Escherichia coli with greater frequency than they would based on chance alone. Network motifs have also been analyzed in the PPI network of Saccharomyces cerevisiae, and from this, it has been concluded that motif topology is correlated with the conservation of motif proteins; besides this, it has been inferred that motifs probably represent the evolutionary-conserved topological units of cellular networks .
True that the representation of a gene regulatory network where the transcriptional regulator is considered the master and the regulated gene the slave [18, 19, 22–24] is appropriate, but such a dichotomous (master-slave) representation cannot be used for a PPI since proteins have multiple functions. In this study, we labeled the protein nodes on the basis of their functional attributes of Gene Ontology (GO) . We then pursued any recurring patterns of the functional attributes of protein interactions. With this new representation of a PPI network and by categorizing the repertoire of network motifs of the same topology into "motif modes," with each motif mode featuring a special topological combination of molecular functions, we have been able to move one step ahead of what was accomplished by Barabási's group . This, in turn, led us to the finding that the evolutionary constraints on the motifs of the same topology are certainly not the same when considering their functional attributes; on the contrary, they vary a great deal.
In that motif modes allow for an analysis that extends beyond motif topology, we were able to look at the properties of motifs of the same topology at a higher resolution than had ever been done before .
Annotation transfer within a protein interaction network
Since we were interested in finding the evolutionary trend of the proteins contained in a motif mode, we built an orthology gene list among the genes of the six organisms (Saccharomyces cerevisiae, Arabidopsis thaliana, Caenorhabditis elegans, Drosophila melanogaster, Mus musculus, and Homo sapiens). When the orthology gene list was used to compute the evolutionary constraint of motifs of different motif topology and motif modes, it was assumed that orthologous proteins also interact with each other. It has been shown that the reliability of the annotation of protein interactions from one organism to another depends on the level of sequence identity for the two pairs of interacting proteins . The joint E-value (the geometric means of the BLAST E-values for the two pairs of interacting proteins) smaller than <10-70 was used as the threshold to evaluate whether such a transfer of the annotation of interactions is reliable . In order to investigate the validity of the orthology gene list by which the interaction annotations were transferred between organisms, we conducted a sequence comparison between orthologs of yeast and five other species. We found 61% of the orthologous protein pairs have an E-value lower than 10-70 (see additional file 2-Sup2.doc, Figure S1). About 40% of the annotation transfers are less reliable; just the same, we have to emphasize that the orthology approach still remains important in terms of decoding many important biological phenomena [22, 27].
Different evolutionary constraints on the motifs of the same topology
A motif mode by chance alone?
As a unique combination of molecular function descriptions of GO, a motif mode could possibly be the result of a special classification that occurs by chance. We randomized the GO annotations (molecular function) on the nodes and re-calculated three features of each motif mode: the motif counts, the conservation fraction and the conservation ratio. We found a significant difference in the motif counts and the conservation ratio of the motif modes with and without the randomization of the GO annotations (p <= 0.0001; see additional file 1-Table S6). Therefore, the motif modes categorized by the current GO annotations and the properties derived from the motif modes (e.g., the level of differential evolutionary conservation) do indeed bear a greater significance than would normally be expected by chance alone.
We used nearly 400 and 650 GO molecular function descriptions at different depths to annotate nearly 5,000 protein nodes in the Saccharomyces cerevisiae PPI network. We employed a motif mode to represent any probable combination of GO annotations with a three-node and four-node topology. To this effect, we collected all of the existing million motif modes and examined the level of evolutionary constraints on their motif constituents (the conservation ratio). We found two interesting distributions of the properties of the motif modes. The first is the distribution of motif counts a motif mode consists of, P(m)~m -φ , where m is the motif count and φ is the degree exponent ~1.6 (Figure 5). This observation may be related to a recent finding that, in a complex network, the large-scale topological organization (characterized by the degree exponents of the scale-free and hierarchical modularity) and the variable counts of the different motif topologies can define each other [28, 29]. Whether there is a direct correlation between the large-scale topological organization and the motif mode-motif count dependence has yet to be determined. The second is the distribution that characterizes the dependence of the number of motif modes on the conservation ratio of that motif mode (Figure 9, Figure S3). Reportedly, the more interconnected the nodes of a motif are with each other, the more conserved the protein constituents of the motif are . If a motif, rather than a single protein (e.g. hub), represents evolutionary-conserved topological units in the tapestry of a PPI network , our study further shows that motifs belonging to different motif modes of the same topology are not under the same level of evolutionary constraints. Fewer motif modes are under higher evolutionary constraints, and the level of differences spans the order of three.
Differences between this and previous studies
The use of GO annotations has provided us with increased insight into protein interactions through the coloring of protein nodes in the interaction map . In this study, we categorized motifs on the basis of the topological combinations of GO terms (molecular function), and this has resulted in our discovery of motif modes. It should not be overlooked that motifs in biological networks have previously been analyzed [18, 22]. The Alon group, for example, developed Mfinder to calculate the motif counts in gene-regulatory networks . Our algorithm is capable not only of conducting motif counts but also of managing the grouping of motif modes based on GO terms; as mentioned earlier, this analysis, in fact, reaches a higher resolution than any other reported before. We recorded the protein identities and GO terms of each motif using our algorithm for the computation of the conservation ratio. We recorded all motif instances (five million) of the motif modes (one million), and they can be used in future analyses though this requires large computer memory. For this very reason, we focused on the motif modes of three nodes and four nodes in this study.
Until now, the evolutionary constraints have only been reported on the level of motif topology, and this by computing the evolutionary constraints of all motifs of the same motif topology . In this study, for the first time, we have been able to distinguish the evolutionary constraints of different motifs of the same topology. A motif topology contains many motif modes which are under different levels of evolutionary constraints. In proposing that motif modes may represent the evolutionary-conserved topological units of a protein interaction network, we have clearly progressed one step farther beyond what has been accomplished before.
It is possible to re-examine the motifs and motif modes of PPIs of model organisms without relying on the orthology information. One can take several real PPI networks of different organisms and, for each network, look independently for the motif they contain and use the GO annotations to define motif modes, then compare if the same motif modes are used in these different organisms. Apart from this, it would be most enlightening to investigate the temporal patterns of various motif modes when gene expression data are used. This would be just like a PPI network visualized within the context of cell-cycles , hubs categorized as "party hubs," or "date hubs,"  and topological changes in transcription regulatory networks observed under environmental or physiological conditions . Furthermore, it would be most worthwhile enhancing the precision of module-detection methods [34–40] and developing specialized tools designed to search for motifs with similar functional annotations. This would greatly assist biologists interested in mining protein interaction networks. PathBlast , TopNet , MAVisto , and FANMOD  are examples of work in these areas.
We have reported on using the molecular function vocabulary of GO to annotate a yeast protein interaction network. The motif mode, i.e., the special topological combination of the molecular functions of interacting proteins, was extracted from the yeast protein interaction network and employed in the analysis of the protein interaction network at a higher resolution than ever before. The distribution of the motif counts for all the motif modes follows a scale-free like fashion. The differential evolutionary constraints on the million motif modes are an indication that motif modes may very well represent the evolutionary-conserved topological units of the protein interaction network.
The yeast (Saccharomyces cerevisiae) PPI dataset is from the DIP database  (July 2004). There are 15,409 interactions in the original data, and after self-interacting protein interactions were removed, 15,129 interactions composed of 4,738 genes (proteins) remained.
Gene Ontology associations
The molecular function of the open reading frame (ORF) name of each gene is based on GO annotations (July 8, 2004). Of 4,738 genes, 857 GO terms are used to describe the molecular functions of genes. No GO term (molecular function) can be found for 289 open reading frames, and these account for 499 interactions. An ORF-GO correspondence table is provided in the additional file 1-Table S7.
Molecular functions of GO terms at different depths
GO terms are organized in structures referred to as 'directed acyclic graphs,' and each term can be traced to different depths in the hierarchies. The molecular function of the GO terms for each protein in our dataset was traced to depths five and six backward to the root. (The root is located at depth 0; "molecular function", GO:0003674, at depth one; "antioxidant activity", GO:0016209, at depth 2). If the GO term for a protein is located at a depth lower than six in the GO tree, the GO term remains unchanged. Our reason for choosing depths five and six is that, based on the statistics, the average depth of the GO terms in our dataset is six (data not shown). The 857 GO terms used to describe a gene's molecular function were then reduced to 398 and 648 terms at depths five and six, respectively.
Algorithm to detect motif modes
We considered a PPI network as a graph and used the method of an adjacency list to represent the graph. In this representation, we labeled the nodes, and for each node, we created a linked list to record its immediate neighbors. To count the occurrences of each motif mode, we modified and employed the graph-searching algorithm, Depth-First-Search (DFS). To start the search, we used each graph node sequentially as the initial node, with a depth limit corresponding to the number of motif nodes. We performed the DFS-based search in a recursive way and implemented a checking procedure to ensure that we did not repeatedly count any identical structure during the search. When we recognized a special topological structure, we proceeded to map the nodes of that structure to those of the motif based on the links through which the nodes were connected to each other. We then recorded the identity of the nodes of the newly found structure and added them to the set of this specific motif mode.
We retrieved the orthologous sequence pairs from InParanoid  (Version 3.0; updated 15, August 2004) for a bootstrap value of 100% and a score of 1.00 in each cluster. We found 1,247 genes of Saccharomyces cerevisiae that have high-confidence orthologs with 1,440 genes from Arabidopsis thaliana, 1,286 genes from Caenorhabditis elegans, 1,397 genes from Drosophila melanogaster, 1,501 genes from Mus musculus, and 1,439 genes from Homo sapiens. We built an orthology gene list using 1,247 genes of Saccharomyces cerevisiae for the protein sequences of the genes orthologous to the five species. Out of 15,129 interactions, 1,247 genes from yeast constitute 2,629 protein interactions.
Conservation fraction and conservation ratio
We then had to find all the constituents of the evolutionary fully conserved motif in the orthology gene list. For each motif topology and each motif mode, we computed the "conservation fraction", i.e., the counts of the evolutionary fully conserved motifs divided by the total motif counts of the motif topology.
To identify the degree to which a motif is under higher evolutionary constraints than would normally be expected by chance alone, we computed "conservation ratio". We randomly assigned the orthology data to the proteins and re-calculated the conservation fraction, as stated above. We repeated this procedure twenty times and obtained an average value which we termed the "random conservation fraction" (see additional file 1-Table S3). The "conservation ratio" of a motif topology or a motif mode is the ratio of the conservation fraction over the "random conservation fraction" of topology or mode, respectively. If the "random conservation fraction" was zero, we identified the "conservation ratio" as non-determined value (NDV).
We thank two reviewers of this publication for their valuable comments and suggestions which helped us greatly improve the revised manuscript. We are also grateful for discussions with Mr. Tzung-Yan Tzou, technical assistance from Mr. Wei-Mo Lee and encouragements from MedicoGenomic Research Center, Kaohsiung Medical University. This work has been supported by the National Science Council of Taiwan (Grant Nos. N92-2113-M-019-001, NSC 93-2113-M-019-001, NSC 94-2213-E-390-002, and NSC 94-2118-M-019-001).
- Giot L, Bader JS, Brouwer C, Chaudhuri A, Kuang B, Li Y, Hao YL, Ooi CE, Godwin B, Vitols E, Vijayadamodar G, Pochart P, Machineni H, Welsh M, Kong Y, Zerhusen B, Malcolm R, Varrone Z, Collis A, Minto M, Burgess S, McDaniel L, Stimpson E, Spriggs F, Williams J, Neurath K, Ioime N, Agee M, Voss E, Furtak K, Renzulli R, Aanensen N, Carrolla S, Bickelhaupt E, Lazovatsky Y, DaSilva A, Zhong J, Stanyon CA, Finley RLJR, White KP, Braverman M, Jarvie T, Gold S, Leach M, Knight J, Shimkets RA, McKenna MP, Chant J, Rothberg JMA: A protein interaction map of Drosophila melanogaster. Science. 2003, 302: 1727-1736. 10.1126/science.1090289.PubMedView ArticleGoogle Scholar
- Schwikowski B, Uetz P, Fields S: A network of protein-protein interactions in yeast. Nat Biotechnol. 2000, 18: 1257-1261. 10.1038/82360.PubMedView ArticleGoogle Scholar
- Uetz P, Giot L, Cagney G, Mansfield TA, Judson RS, Knight JR, Lockshon D, Narayan V, Srinivasan M, Pochart P, Qureshi-Emili A, Li Y, Godwin B, Conover D, Kalbfleisch T, Vijayadamodar G, Yang M, Johnston M, Fields S, Rothberg JM: A comprehensive analysis of protein-protein interactions in Saccharomyces cerevisiae. Nature. 2000, 403: 623-627. 10.1038/35001009.PubMedView ArticleGoogle Scholar
- Walhout AJ, Boulton SJ, Vidal M: Yeast two-hybrid systems and protein interaction mapping projects for yeast and worm. Yeast. 2000, 17: 88-94. 10.1002/1097-0061(20000630)17:2<88::AID-YEA20>3.0.CO;2-Y.PubMedPubMed CentralView ArticleGoogle Scholar
- Xenarios I, Eisenberg D: Protein interaction databases. Curr Opin Biotechnol. 2001, 12: 334-339. 10.1016/S0958-1669(00)00224-X.PubMedView ArticleGoogle Scholar
- Barabasi AL, Oltvai ZN: Network biology: understanding the cell's functional organization. Nat Rev Genet. 2004, 5: 101-113. 10.1038/nrg1272.PubMedView ArticleGoogle Scholar
- Jeong H, Mason SP, Barabasi AL, Oltvai ZN: Lethality and centrality in protein networks. Nature. 2001, 411: 41-42. 10.1038/35075138.PubMedView ArticleGoogle Scholar
- Wagner A: The yeast protein interaction network evolves rapidly and contains few redundant duplicate genes. Mol Biol Evol. 2001, 18: 1283-1292.PubMedView ArticleGoogle Scholar
- Yook SH, Oltvai ZN, Barabasi AL: Functional and topological characterization of protein interaction networks. Proteomics. 2004, 4: 928-942. 10.1002/pmic.200300636.PubMedView ArticleGoogle Scholar
- Wuchty S: Interaction and domain networks of yeast. Proteomics. 2002, 2: 1715-1723. 10.1002/1615-9861(200212)2:12<1715::AID-PROT1715>3.0.CO;2-O.PubMedView ArticleGoogle Scholar
- Qin H, Lu HH, Wu WB, Li WH: Evolution of the yeast protein interaction network. Proc Natl Acad Sci U S A. 2003, 100: 12820-12824. 10.1073/pnas.2235584100.PubMedPubMed CentralView ArticleGoogle Scholar
- Wagner A: Does selection mold molecular networks?. Sci STKE. 2003, 2003: E41-Google Scholar
- Albert R, Jeong H, Barabasi AL: Error and attack tolerance of complex networks. Nature. 2000, 406: 378-382. 10.1038/35019019.PubMedView ArticleGoogle Scholar
- Fraser HB, Wall DP, Hirsh AE: A simple dependence between protein evolution rate and the number of protein-protein interactions. BMC Evol Biol. 2003, 3: 11-10.1186/1471-2148-3-11.PubMedPubMed CentralView ArticleGoogle Scholar
- Fraser HB, Hirsh AE, Steinmetz LM, Scharfe C, Feldman MW: Evolutionary rate in the protein interaction network. Science. 2002, 296: 750-752. 10.1126/science.1068696.PubMedView ArticleGoogle Scholar
- Hahn MW, Conant GC, Wagner A: Molecular evolution in large genetic networks: does connectivity equal constraint?. J Mol Evol. 2004, 58: 203-211. 10.1007/s00239-003-2544-0.PubMedView ArticleGoogle Scholar
- Guelzim N, Bottani S, Bourgine P, Kepes F: Topological and causal structure of the yeast transcriptional regulatory network. Nat Genet. 2002, 31: 60-63. 10.1038/ng873.PubMedView ArticleGoogle Scholar
- Milo R, Shen-Orr S, Itzkovitz S, Kashtan N, Chklovskii D, Alon U: Network motifs: simple building blocks of complex networks. Science. 2002, 298: 824-827. 10.1126/science.298.5594.824.PubMedView ArticleGoogle Scholar
- Shen-Orr SS, Milo R, Mangan S, Alon U: Network motifs in the transcriptional regulation network of Escherichia coli. Nat Genet. 2002, 31: 64-68. 10.1038/ng881.PubMedView ArticleGoogle Scholar
- Wall ME, Dunlop MJ, Hlavacek WS: Multiple functions of a feed-forward-loop gene circuit. J Mol Biol. 2005, 349: 501-514. 10.1016/j.jmb.2005.04.022.PubMedView ArticleGoogle Scholar
- Lee TI, Rinaldi NJ, Robert F, Odom DT, Bar-Joseph Z, Gerber GK, Hannett NM, Harbison CT, Thompson CM, Simon I, Zeitlinger J, Jennings EG, Murray HL, Gordon DB, Ren B, Wyrick JJ, Tagne JB, Volkert TL, Fraenkel E, Gifford DK, Young RA: Transcriptional regulatory networks in Saccharomyces cerevisiae. Science. 2002, 298: 799-804. 10.1126/science.1075090.PubMedView ArticleGoogle Scholar
- Wuchty S, Oltvai ZN, Barabasi AL: Evolutionary conservation of motif constituents in the yeast protein interaction network. Nat Genet. 2003, 35: 176-179. 10.1038/ng1242.PubMedView ArticleGoogle Scholar
- Yeger-Lotem E, Sattath S, Kashtan N, Itzkovitz S, Milo R, Pinter RY, Alon U, Margalit H: Network motifs in integrated cellular networks of transcription-regulation and protein-protein interaction. Proc Natl Acad Sci U S A. 2004, 101: 5934-5939. 10.1073/pnas.0306752101.PubMedPubMed CentralView ArticleGoogle Scholar
- Mangan S, Alon U: Structure and function of the feed-forward loop network motif. Proc Natl Acad Sci U S A. 2003, 100: 11980-11985. 10.1073/pnas.2133841100.PubMedPubMed CentralView ArticleGoogle Scholar
- GO Database. [http://www.geneontology.org/]
- Yu H, Luscombe NM, Lu HX, Zhu X, Xia Y, Han JD, Bertin N, Chung S, Vidal M, Gerstein M: Annotation transfer between genomes: protein-protein interologs and protein-DNA regulogs. Genome Res. 2004, 14: 1107-1118. 10.1101/gr.1774904.PubMedPubMed CentralView ArticleGoogle Scholar
- Wuchty S, Barabasi AL, Ferdig MT: Stable evolutionary signal in a Yeast protein interaction network. BMC Evol Biol. 2006, 6: 8-10.1186/1471-2148-6-8.PubMedPubMed CentralView ArticleGoogle Scholar
- Vazquez A, Dobrin R, Sergi D, Eckmann JP, Oltvai ZN, Barabasi AL: The topological relationship between the large-scale attributes and local interaction patterns of complex networks. Proc Natl Acad Sci U S A. 2004, 101: 17940-17945. 10.1073/pnas.0406024101.PubMedPubMed CentralView ArticleGoogle Scholar
- Vazquez A, Oliveira JG, Barabasi AL: Inhomogeneous evolution of subgraphs and cycles in complex networks. Phys Rev E Stat Nonlin Soft Matter Phys. 2005, 71: 025103-PubMedView ArticleGoogle Scholar
- Breitkreutz BJ, Stark C, Tyers M: Osprey: a network visualization system. Genome Biol. 2003, 4: R22-10.1186/gb-2003-4-3-r22.PubMedPubMed CentralView ArticleGoogle Scholar
- de Lichtenberg U, Jensen LJ, Brunak S, Bork P: Dynamic complex formation during the yeast cell cycle. Science. 2005, 307: 724-727. 10.1126/science.1105103.PubMedView ArticleGoogle Scholar
- Han JD, Bertin N, Hao T, Goldberg DS, Berriz GF, Zhang LV, Dupuy D, Walhout AJ, Cusick ME, Roth FP, Vidal M: Evidence for dynamically organized modularity in the yeast protein-protein interaction network. Nature. 2004, 430: 88-93. 10.1038/nature02555.PubMedView ArticleGoogle Scholar
- Luscombe NM, Babu MM, Yu H, Snyder M, Teichmann SA, Gerstein M: Genomic analysis of regulatory network dynamics reveals large topological changes. Nature. 2004, 431: 308-312. 10.1038/nature02782.PubMedView ArticleGoogle Scholar
- Dunn R, Dudbridge F, Sanderson CM: The use of edge-betweenness clustering to investigate biological function in protein interaction networks. BMC Bioinformatics. 2005, 6: 39-10.1186/1471-2105-6-39.PubMedPubMed CentralView ArticleGoogle Scholar
- Rives AW, Galitski T: Modular organization of cellular networks. Proc Natl Acad Sci U S A. 2003, 100: 1128-1133. 10.1073/pnas.0237338100.PubMedPubMed CentralView ArticleGoogle Scholar
- Pereira-Leal JB, Enright AJ, Ouzounis CA: Detection of functional modules from protein interaction networks. Proteins. 2004, 54: 49-57. 10.1002/prot.10505.PubMedView ArticleGoogle Scholar
- Ma HW, Zhao XM, Yuan YJ, Zeng AP: Decomposition of metabolic network into functional modules based on the global connectivity structure of reaction graph. Bioinformatics. 2004, 20: 1870-1876. 10.1093/bioinformatics/bth167.PubMedView ArticleGoogle Scholar
- Poyatos JF, Hurst LD: How biologically relevant are interaction-based modules in protein networks?. Genome Biol. 2004, 5: R93-10.1186/gb-2004-5-11-r93.PubMedPubMed CentralView ArticleGoogle Scholar
- Spirin V, Mirny LA: Protein complexes and functional modules in molecular networks. Proc Natl Acad Sci U S A. 2003, 100: 12123-12128. 10.1073/pnas.2032324100.PubMedPubMed CentralView ArticleGoogle Scholar
- Zhang LV, King OD, Wong SL, Goldberg DS, Tong AH, Lesage G, Andrews B, Bussey H, Boone C, Roth FP: Motifs, themes and thematic maps of an integrated Saccharomyces cerevisiae interaction network. J Biol. 2005, 4: 6-10.1186/jbiol23.PubMedPubMed CentralView ArticleGoogle Scholar
- Kelley BP, Yuan B, Lewitter F, Sharan R, Stockwell BR, Ideker T: PathBLAST: a tool for alignment of protein interaction networks. Nucleic Acids Res. 2004, 32: W83-W88. 10.1093/nar/gnh080.PubMedPubMed CentralView ArticleGoogle Scholar
- Yu H, Zhu X, Greenbaum D, Karro J, Gerstein M: TopNet: a tool for comparing biological sub-networks, correlating protein properties with topological statistics. Nucleic Acids Res. 2004, 32: 328-337. 10.1093/nar/gkh164.PubMedPubMed CentralView ArticleGoogle Scholar
- Schreiber F, Schwobbermeyer H: MAVisto: a tool for the exploration of network motifs. Bioinformatics. 2005, 21: 3572-3574. 10.1093/bioinformatics/bti556.PubMedView ArticleGoogle Scholar
- Wernicke S, Rasche F: FANMOD: a tool for fast network motif detection. Bioinformatics. 2006Google Scholar
- DIP database. [http://dip.doe-mbi.ucla.edu/]
- InParanoid. [http://inparanoid.cgb.ki.se/index.html]