ࡱ > C C B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B q` < bjbjqPqP 4 : : 3 H H H H X X X X b x N = L $ h B 1 1 1 H H B 1 r H 8 1 6
_ X X G T b x 8 " 1 1 1 1 1 1 1 s 1 1 1 b 1 1 1 1 P
Q [ dJ P
[ H H H H H H Doc-file of the Webtutorial at: HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/" http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/
Tutorial: The PluriNetWork / ExprEssence framework
Contents
HYPERLINK \l "Motivation" Motivation
HYPERLINK \l "The_PluriNetWork" The PluriNetWork
HYPERLINK \l "Exploring_the_PluriNetWork" Exploring the PluriNetWork
HYPERLINK \l "Finding_a_gene" Finding a gene in the PluriNetWork
HYPERLINK \l "Inspecting" Inspecting Attribute Values
HYPERLINK \l "Getting" Getting the underlying Pubmed article(s)
HYPERLINK \l "bingo" BiNGO Gene Ontology analysis
HYPERLINK \l "ExprEssence" ExprEssence
HYPERLINK \l "Installation" Installation
HYPERLINK \l "General_ExprEssence_Workflow" General ExprEssence Workflow
HYPERLINK \l "CaseStudies" Case studies
HYPERLINK \l "MEF_pips_ips" MEF to piPS and piPS to iPS
HYPERLINK \l "ESC_to_Epiblast" Embryonic Stem Cell to Epiblast Stem Cell state
HYPERLINK \l "NodeColor_v" NodeColor- HYPERLINK \l "NodeColor_Oct4_CaseStudy" Oct4 knockdown data
HYPERLINK \l "References" References
HYPERLINK \l "About" About
Motivation
We wish to obtain insights into biological processes by building up process- and species-specific protein/gene interaction and regulation networks and combining them with differential biological data (e.g. gene expression data for two time points). Towards this aim, we built up a network describing pluripotency in mouse and used it for stem cell data analysis.
The PluriNetWork
The HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/MEF_piPS_iPS.cys" PluriNetWork, described in the HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/PluriNetWork.pdf" PluriNetWork paper, is a manually curated protein/gene interaction and regulation network with the purpose of describing pluripotency in mouse. The structure of the PluriNetWork is straightforward; each node represents one gene and its corresponding protein, and links are either interactions, stimulations or inhibitions. As of July 2010, the network is based upon 177 publications, consists of 274 nodes (genes/proteins) and 574 edges (links) - each representing a direct interaction or regulation between two nodes. The PluriNetWork is explored, maintained and analyzed using Cytoscape (see Figure 1), (downloadable at HYPERLINK "http://www.cytoscape.org/download.html" http://www.cytoscape.org/download.html).
Figure SEQ Fig. \* ARABIC 1. The PluriNetWork, displayed using the Cytoscape network visualization and data integration platform.
Exploring the PluriNetWork
Apart from the circuit-like layout, the PluriNetWork data is represented by two tables - one table describing the nodes (genes/proteins) and their attributes and the other the edges (links) and their attributes. At any given time, only one table can be shown in the Cytoscape Data Panel. To switch between the tables, you can click at the bottom of the Data Panel on the tab Node Attribute Browser or Edge Attribute Browser (see Figure 2).
Figure SEQ Fig. \* ARABIC 2. The Data Panel in Cytoscape.
Finding a gene in the PluriNetWork
Click on the button right to the search input field to configure your search options (see Figure 3). The radio button that appears in the configuration dialog must be set to Nodes.
Figure SEQ Fig. \* ARABIC 3. Starting the search configuration dialog.
Select an attribute containing the values you will be using for the search (e.g. an identifier such as MGI Symbol, to search for Pou5f1, Sox2, Nanog, Esrrb, etc).
Type your query into the search input field.
If there exists a node (gene/protein) with the attribute/value pair used for the search, this node will be moved to the center of the main panel (if the value cannot be found, the search input field will be marked by a red background).
A subnetwork can easily be retrieved in Cytoscape by selecting the desired nodes and edges and clicking on File->New->Network->From selected nodes and edges.
Inspecting Attribute Values
In the PluriNetWork, for each node (gene/protein) and edge (link), attributes can be browsed as follows.
Mark the node/edge of interest by a left mouse click (the color should change); if more than one node/edge shall be selected, hold Shift during selection, or select all nodes/edges in an area defined using the left mouse button.
Select the Node Attribute Browser or Edge Attribute Browser at the bottom of the Data Panel.
To define which attributes shall be shown, click on the first button of the Data Panel (see Figure 4), and then tick the corresponding checkboxes in the dialog window that opens up.
The selected attribute values of the selected nodes/edges are then shown in the Data Panel.
Figure SEQ Fig. \* ARABIC 4. Starting the attribute browser.
Getting the underlying Pubmed article(s)
For each link, at least one reference is available in the PluriNetWork to back it up. It can be retrieved as follows.
Mark the edge for which you want to get the PubMed entry.
Choose the tab of the Edge Attribute Browser at the bottom of the Data Panel.
Left-click on the Pubmed ID attribute (if necessary, follow the section HYPERLINK \l "Inspecting" Inspecting Attribute Values, above, so that the Pubmed ID attribute is shown.)
The standard browser should open with the abstract of the Pubmed article. More references may be available by inspecting the attributes Pubmed ID Source 2, etc. (If necessary, follow the section HYPERLINK \l "Inspecting" Inspecting Attribute Values, above, so that these attributes are shown.)
BINGO Gene Ontology analysis
Cytoscape can be enhanced by a variety of plugins, allowing, for example, to perform network statistics, or to import data of different formats. For the PluriNetWork paper, we performed an overrepresentation analysis of the biological processes and molecular functions, based on the Gene Ontology Terms of all 274 genes. We used BiNGO ADDIN EN.CITE Ashburner200014514514517Ashburner, M.Ball, C. A.Blake, J. A.Botstein, D.Butler, H.Cherry, J. M.Davis, A. P.Dolinski, K.Dwight, S. S.Eppig, J. T.Harris, M. A.Hill, D. P.Issel-Tarver, L.Kasarskis, A.Lewis, S.Matese, J. C.Richardson, J. E.Ringwald, M.Rubin, G. M.Sherlock, G.Department of Genetics, Stanford University School of Medicine, California, USA. cherry@stanford.eduGene ontology: tool for the unification of biology. The Gene Ontology ConsortiumNat GenetNat Genet25-9251AnimalsComputer Communication NetworksDatabases, FactualEukaryotic Cells/*physiology*GenesHumansMetaphysicsMiceMolecular Biology/*trends*Sequence Analysis, DNA*Terminology as Topic2000May10802651http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=10802651[1] with the GO Slim Generic Gene Ontology Annotation ADDIN EN.CITE Camon200414414414417Camon, E.Magrane, M.Barrell, D.Lee, V.Dimmer, E.Maslen, J.Binns, D.Harte, N.Lopez, R.Apweiler, R.European Bioinformatics Institute (EBI), Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SD, UK. goa@ebi.ac.ukThe Gene Ontology Annotation (GOA) Database: sharing knowledge in Uniprot with Gene OntologyNucleic Acids ResNucleic Acids ResD262-632Database issueAnimalsComputational Biology*Databases, Genetic*Databases, Protein*GenesHumansInformation Storage and RetrievalInternetProteome/chemistry/genetics/metabolismProteomics*Terminology as Topic2004Jan 114681408http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=14681408[2], which is a set of high-level GO terms. The significance level was set to p=0.05 (hypergeometric test, Benjamini & Hochberg False Discovery Rate (FDR) correction). Overrepresented terms are visualized in Figure 5 below.
To perform this analysis yourself, start Cytoscape, using the Plugins->Manage Plugins dialog to install the BiNGO plugin available in the Functional Enrichment folder, and open the file HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/MEF_piPS_iPS.cys" MEF_PIPS_IPS.cys, which includes the PluriNetWork. Select all nodes in the network, using Select->Nodes->Select All nodes in the main Cytoscape menu. In the Data Panel of Cytoscape, select the Node Attribute Browser, and, following the instructions on HYPERLINK \l "Inspecting" Inspecting Attribute Values, select all values of the attribute MGI Symbol and copy the values into the clipboard. Then, start BiNGO via the Plugins menu. Define a name for your BiNGO analysis and select Paste Genes from Text. Then, paste the contents of your clipboard into the open dialog field. In the field Select organism/annotation, please select Mus musculus. For the analysis described here, all other parameters are default values. Start the analysis by clicking on Start BiNGO. A new network is generated with the name you have chosen above. This network consists of the GO Terms which could be mapped to gene names in the starting network. GO Terms which are overrepresented in your network will be highlighted. In the panel BiNGO output, you will find more information about the GO Terms (GO identifiers, significance of overrepresentation, names of genes mapped to this GO Term).
Figure SEQ Fig. \* ARABIC 5. Result of GO analysis by BINGO.
In Figure 5, enriched cellular components are the nuclear chromosome and the nucleoplasm, reflecting the prominence of transcription factors and epigenetic factors in the PluriNetWork. Biological Process Terms are enriched for transcription, DNA metabolism, protein modification, cell differentiation / cell death, embryonic development, cell cycle, epigenetic regulation of gene expression and signal transduction. Enriched molecular functions such as transcription factor activity, chromatin binding, receptor binding and protein kinase activities fit well to the enriched biological processes.
ExprEssence - Condensing Networks
ExprEssence is a Cytoscape plugin analyzing a network of genes/proteins together with differential biological data, as measured in an experiment. It highlights the links across which the largest amount of change can be observed, given two experimental data sets. More specifically, ExprEssence condenses networks so that they contain only those links between genes/proteins, along which a large amount of change in (expression) values takes place. These links are called most differentially altered. The percentage of most differentially altered links to be highlighted can be set by the user, using a slider. Highlighting identifies hypotheses about the startup or the shutdown of interactions, stimulations and inhibitions.
Installation
The up-to-date ExprEssence.jar file can be downloaded at HYPERLINK "http://sourceforge.net/projects/expressence/" sourceforge.net.
Move the file to the plugins folder in the main Cytoscape folder, and (re-)start Cytoscape.
General ExprEssence Workflow
The following is the general ExprEssence workflow.
Open Cytoscape and open a Cytoscape session file (a cys file such as HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/MEF_piPS_iPS.cys" MEF_PIPS_IPS.cys). If not already done, import your own gene expression data (or other high-throughput data, as described HYPERLINK "http://cytoscape.wodaklab.org/wiki/Cytoscape_User_Manual/Attributes" \l "Import_Attribute_Table_Files" here and in ADDIN EN.CITE ADDIN EN.CITE.DATA [3]). Data are already imported for the case studies below. In the Network tab of the Control Panel on the left, select the network you wish to condense.
Click the condense! button in the tool bar (see Figure 6).
Figure SEQ Fig. \* ARABIC 6. The condense button.
Choose the attribute-containing data for the first experiment on the left (such as GSE14012_MEF_mean; see Figure 7 below: arrow 1) and the attribute-containing data for the second experiment on the right (such as GSE14012_piPS_mean; see Figure 7 below: arrow 2).
Figure SEQ Fig. \* ARABIC 7. Window for selecting attributes in ExprEssence.
Press OK (Figure 7: arrow 3).
If you cannot provide variance data, click No variance data (see Figure 8).If you can provide variance data (such as GSE14012_MEF_var and GSE14012_piPS_var), select these as in step 3 and enter the number of replicates (here, for GSE14012_MEF_var the number is 4 and for GSE14012_piPS_var the number is 3, as in Figure 8). Then press OK.
Figure SEQ Fig. \* ARABIC 8. Window for selecting the variance attributes in ExprEssence.
A new network will then be generated and the Control Panel of Cytoscape (on the left side of the Cytoscape window) displays a slider to adjust the percentage P of most differentially altered links (see Figure 9), allowing to vary the degree of network condensation.
Figure SEQ Fig. \* ARABIC 9. ExprEssence Slider.
The percentage P can be adjusted for both startups (red, top part of slider), and shutdowns (green, bottom part of slider). To keep the P % most differentially altered shutdowns, select a lower threshold of P %. To keep the P % most differentially altered startups, select an upper threshold of 100 P %.
In tabular form, an ExprEssence workflow asks you to provide the following data, with respect to the steps of the General ExprEssence Workflow:
StepsTo doExampleFile: (Step 1)Select a Cytoscape file HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/MEF_piPS_iPS.cys" MEF_PIPS_IPS.cysAttribute: (Step 3)
mean valuesSelect mean values for the first (left) and the second (right) condition.left: GSE14012_MEF_mean
right: GSE14012_piPS_meanAttribute: (Step 4)
variances (replicates)
[optional] Optional: Select variance data for the first (left) and the second (right) condition and provide the number of replicates.left: GSE14012_MEF_var (replicates: 4)
right: GSE14012_piPS_var (replicates: 3)Slider: (Step 6)
lower [optional]Optional: Adjust the lower slider so that P % of the shutdowns are shown.10%Slider: (Step 6)
upper [optional]Optional: Adjust the upper slider so that 100-P % of the shutdowns are shown.90%
Table SEQ Table \* ARABIC 1. Basic ExprEssence user input using the PluriNetWork.
Finally, the following Table describes the general workflow of an ExprEssence analysis using the PluriNetWork:
Input: Gene expression measurements describing an experiment. We always compare two measurements, for example before and after, or condition 1 and condition 2.Analysis:ExprEssence highlighting of the top startups and shutdowns of interactions, stimulations and inhibitions in the PluriNetWork.Output: Hypotheses for (regulatory) mechanisms acting in the course of the experiment, e.g. highlighting the startup of a stimulation, if the expression of both stimulator and target go up.Interpretation: Plausibility check / confirmation experiments for the (regulatory) mechanisms found.
Table SEQ Table \* ARABIC 2. General workflow of an ExprEssence analysis using the PluriNetWork.
Case studies
In the following case studies we used ExprEssence, combining the PluriNetWork with expression data and highlighting the most significant changes of gene expression along links connecting genes/proteins of the PluriNetWork.
MEF to piPS and piPS to iPS
Transition from mouse embryonic fibroblast (MEF) to partially induced pluripotent stem cells (piPS) and from piPS to induced pluripotent stem cells (iPS).
As described in the PluriNetWork paper, we used microarray data by Sridharan et al. ADDIN EN.CITE Sridharan200915915915917Sridharan, R.Tchieu, J.Mason, M. J.Yachechko, R.Kuoy, E.Horvath, S.Zhou, Q.Plath, K.David Geffen School of Medicine, University of California, Los Angeles, CA 90095, USA.Role of the murine reprogramming factors in the induction of pluripotencyCellCell364-771362AnimalsCell DifferentiationDNA-Binding Proteins/metabolismEmbryonic Stem Cells/*cytologyFibroblasts/cytologyMice/*metabolismNuclear Proteins/metabolism*Nuclear ReprogrammingPluripotent Stem Cells/*cytology2009Jan 2319167336http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=19167336[4] to investigate the induction of pluripotency.
Input: Microarray data describing (1) fibroblasts, (2) partially induced and (3) fully induced pluripotent stem cells. We first compare (1) and (2), and then (2) and (3).Analysis:ExprEssence highlighting of the top 10% startups and the top 10% shutdowns.Output: Hypotheses for (regulatory) mechanisms acting in the course of the induction of pluripotency.Interpretation: Comparison of our results with Sridharan et al ADDIN EN.CITE Sridharan200915915915917Sridharan, R.Tchieu, J.Mason, M. J.Yachechko, R.Kuoy, E.Horvath, S.Zhou, Q.Plath, K.David Geffen School of Medicine, University of California, Los Angeles, CA 90095, USA.Role of the murine reprogramming factors in the induction of pluripotencyCellCell364-771362AnimalsCell DifferentiationDNA-Binding Proteins/metabolismEmbryonic Stem Cells/*cytologyFibroblasts/cytologyMice/*metabolismNuclear Proteins/metabolism*Nuclear ReprogrammingPluripotent Stem Cells/*cytology2009Jan 2319167336http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=19167336[4].
Table SEQ Table \* ARABIC 3. Workflow of the ExprEssence analysis of the induction of pluripotency.
The file HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/MEF_piPS_iPS.cys" MEF_PIPS_IPS.cys provides the PluriNetWork, gene expression data (plus variance data) mapped onto the PluriNetWork, as well as the already condensed networks, which we will interpret after describing how to produce them. Before integrating the gene expression data of Sridharan et al ADDIN EN.CITE Sridharan200915915915917Sridharan, R.Tchieu, J.Mason, M. J.Yachechko, R.Kuoy, E.Horvath, S.Zhou, Q.Plath, K.David Geffen School of Medicine, University of California, Los Angeles, CA 90095, USA.Role of the murine reprogramming factors in the induction of pluripotencyCellCell364-771362AnimalsCell DifferentiationDNA-Binding Proteins/metabolismEmbryonic Stem Cells/*cytologyFibroblasts/cytologyMice/*metabolismNuclear Proteins/metabolism*Nuclear ReprogrammingPluripotent Stem Cells/*cytology2009Jan 2319167336http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=19167336[4] (GSE14012), we normalized them (using the "rma" method from the affy package of Bioconductor ADDIN EN.CITE Irizarry200327227227217Irizarry, R. A.Hobbs, B.Collin, F.Beazer-Barclay, Y. D.Antonellis, K. J.Scherf, U.Speed, T. P.Department of Biostatistics, Johns Hopkins University, Baltimore, MD 21205, USA. rafa@jhu.eduExploration, normalization, and summaries of high density oligonucleotide array probe level dataBiostatisticsBiostatistics249-6442AlgorithmsAnimalsDNA Probes/*genetics*Data Interpretation, StatisticalGene Expression Profiling/statistics & numerical dataHumansLinear ModelsMiceNormal DistributionOligonucleotide Array Sequence Analysis/*methodsReproducibility of ResultsStatistics, Nonparametric2003Apr12925520http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=12925520[5]). As a result, nodes (genes) in the PluriNetWork now have the following attributes:
(1) GSE14012_MEF_mean and GSE14012_MEF_var: Mean and variance of gene expression in fibroblasts (MEF, mouse embryonic fibroblasts).
(2) GSE14012_piPS_mean and GSE14012_piPS_var: Mean and variance of gene expression in partially induced pluripotent stem cells (piPS).
(3) GSE14012_iPS_mean and GSE14012_iPS_var: Mean and variance of gene expression in induced pluripotent stem cells (iPS).
Transition from MEF to piPS.
ExprEssence is now used to derive the first condensed network comparing (1) MEF and (2) piPS. In the General ExprEssence Workflow HYPERLINK \l "GEW" above, the following data are provided, resulting in Figure 10 below:
Figure
Steps
10File: HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/MEF_piPS_iPS.cys" MEF_PIPS_IPS.cysAttribute:
mean valuesleft: GSE14012_MEF_mean
right: GSE14012_piPS_meanAttribute:
variances (replicates)left: GSE14012_MEF_var (replicates: 4)
right: GSE14012_piPS_var (replicates: 3)Slider:
lower10%Slider:
upper90%
Table SEQ Table \* ARABIC 4. User input for reproduction of Figure 10.
Figure SEQ Fig. \* ARABIC 10. Transition from MEF to piPS.
To interpret the figure, we first describe some background on the experiment. According to Sridharan et al. ADDIN EN.CITE Sridharan200915915915917Sridharan, R.Tchieu, J.Mason, M. J.Yachechko, R.Kuoy, E.Horvath, S.Zhou, Q.Plath, K.David Geffen School of Medicine, University of California, Los Angeles, CA 90095, USA.Role of the murine reprogramming factors in the induction of pluripotencyCellCell364-771362AnimalsCell DifferentiationDNA-Binding Proteins/metabolismEmbryonic Stem Cells/*cytologyFibroblasts/cytologyMice/*metabolismNuclear Proteins/metabolism*Nuclear ReprogrammingPluripotent Stem Cells/*cytology2009Jan 2319167336http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=19167336[4], partial induction features changes associated with c-Myc, including metabolic regulation and transcriptional repression of somatic gene expression. Inspection of the PluriNetWork links to/from c-Myc indicates that its role in metabolic regulation is not documented well in the specific literature; Sridharan et al. ADDIN EN.CITE Sridharan200915915915917Sridharan, R.Tchieu, J.Mason, M. J.Yachechko, R.Kuoy, E.Horvath, S.Zhou, Q.Plath, K.David Geffen School of Medicine, University of California, Los Angeles, CA 90095, USA.Role of the murine reprogramming factors in the induction of pluripotencyCellCell364-771362AnimalsCell DifferentiationDNA-Binding Proteins/metabolismEmbryonic Stem Cells/*cytologyFibroblasts/cytologyMice/*metabolismNuclear Proteins/metabolism*Nuclear ReprogrammingPluripotent Stem Cells/*cytology2009Jan 2319167336http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=19167336[4] identify this feature by Gene Ontology analysis of target genes bound by c-Myc, based on ChIP-derived information. The role of c-Myc in transcriptional repression is concordant with the observation that in iPS generation, c-Myc may be substituted by small molecules with a histone deacetylation effect (e.g. valproic acid, VPA) ADDIN EN.CITE Sridharan200915915915917Sridharan, R.Tchieu, J.Mason, M. J.Yachechko, R.Kuoy, E.Horvath, S.Zhou, Q.Plath, K.David Geffen School of Medicine, University of California, Los Angeles, CA 90095, USA.Role of the murine reprogramming factors in the induction of pluripotencyCellCell364-771362AnimalsCell DifferentiationDNA-Binding Proteins/metabolismEmbryonic Stem Cells/*cytologyFibroblasts/cytologyMice/*metabolismNuclear Proteins/metabolism*Nuclear ReprogrammingPluripotent Stem Cells/*cytology2009Jan 2319167336http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=19167336[4], ADDIN EN.CITE Huangfu200818218218217Huangfu, D.Osafune, K.Maehr, R.Guo, W.Eijkelenboom, A.Chen, S.Muhlestein, W.Melton, D. A.Department of Stem Cell and Regenerative Biology, Howard Hughes Medical Institute, Harvard Stem Cell Institute, Harvard University, 7 Divinity Avenue, Cambridge, Massachusetts 02138, USA.Induction of pluripotent stem cells from primary human fibroblasts with only Oct4 and Sox2Nat BiotechnolNat Biotechnol1269-752611AnimalsBiotechnology/methodsCell DifferentiationCells, CulturedCulture MediaEmbryonic Stem Cells/cytologyEpigenesis, GeneticFibroblasts/*cytology/metabolismGene Expression ProfilingHumansMice*Nuclear Reprogramming/drug effectsOctamer Transcription Factor-3/*metabolismPluripotent Stem Cells/*cytology/metabolismSOXB1 Transcription Factors/*metabolismValproic Acid/*pharmacology2008Nov18849973http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=18849973[6]. In the PluriNetWork, histone deacetylation and other epigenetic phenomena are included, but they are not connected to c-Myc. Moreover, we note that c-Myc is downregulated.
Instead, startup of the stimulation of epigenetic effectors is highlighted around Oct4/Pou5f1, based on two recent publications ADDIN EN.CITE ADDIN EN.CITE.DATA [7,8], owing to the strong upregulation of Oct4. Consequently, by ExprEssence condensation of the PluriNetWork highlighting putative mechanisms of partial induction (Figure 10), we find a pronounced startup of epigenetic phenomena, associated with the upregulation of Oct4/Pou5f1. In particular, Oct4 stimulation of the histone deacetylase Hdac2 starts up. Moreover, interactions of Oct4 with Hells, Smarca4, Mta1, Gata2a, Smarca5, Trim28, Dhx9, Hira, Cabin1, SSrp1, Kdm1a, Ctcf, Phc1, Rnf2, and Eed all start up. Also, Oct4/Pou5f1 inhibition by Cdx2 ADDIN EN.CITE Niwa200518318318317Niwa, H.Toyooka, Y.Shimosato, D.Strumpf, D.Takahashi, K.Yagi, R.Rossant, J.Laboratory for Pluripotent Cell Studies, RIKEN Center for Developmental Biology (CDB), 2-2-3 Minatojima-minamimachi, Kobe, Hyogo 650-0047, Japan. niwa@cdb.riken.jpInteraction between Oct3/4 and Cdx2 determines trophectoderm differentiationCellCell917-291235Animals*Blastocyst/cytology/physiologyCell Differentiation/*physiology*Cell LineageCells, CulturedEnzyme ActivationGene Expression Regulation, DevelopmentalHomeodomain Proteins/genetics/*metabolismMiceMice, Inbred C57BLOctamer Transcription Factor-3/genetics/*metabolismOrganic Cation Transport Proteins/genetics/*metabolismPluripotent Stem Cells/cytology/*physiologyRecombinant Fusion Proteins/genetics/metabolismTranscription Factors/genetics/*metabolism2005Dec 216325584http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=16325584[9], by Wwp2 ADDIN EN.CITE Liao18418418417Liao, B.Jin, Y.Key Laboratory of Stem Cell Biology, Institute of Health Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences/Shanghai Jiao Tong University School of Medicine; 225 South Chongqing Road, Shanghai 200025, China.Wwp2 mediates Oct4 ubiquitination and its own auto-ubiquitination in a dosage-dependent mannerCell ResCell Res332-44203AnimalsCell DifferentiationCell Line, TumorDimerizationGene Knockdown TechniquesMiceOctamer Transcription Factor-3/*metabolismRNA InterferenceTretinoin/pharmacologyUbiquitin-Protein Ligases/genetics/*metabolismUbiquitination2010Mar19997087http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=19997087[10] and by some nuclear receptors (Nr2f1, Nr2f2, Nr6a1) is shut down. Around c-Myc, we observe a few shutdowns (of its stimulation by Stat3 and Tcf3, and of its Sox2 stimulation), owing to its downregulation from MEF to piPS cells.
Transition from piPS to iPS.
ExprEssence is now used to derive a second condensed network, comparing (2) piPS and (3) iPS. In the General ExprEssence Workflow HYPERLINK \l "GEW" above, the following data are provided, resulting in Figure 11 below:
Figure
Steps
11File: HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/MEF_piPS_iPS.cys" MEF_PIPS_IPS.cysAttribute:
mean valuesleft: GSE14012_piPS_mean
right: GSE14012_iPS_meanAttribute:
variances (replicates)left: GSE14012_piPS_var (replicates: 3)
right: GSE14012_iPS_var (replicates: 3)Slider:
lower10%Slider:
upper90%
Table SEQ Table \* ARABIC 5. User input for reproduction of Figure 11.
Sridharan et al. ADDIN EN.CITE Sridharan200915915915917Sridharan, R.Tchieu, J.Mason, M. J.Yachechko, R.Kuoy, E.Horvath, S.Zhou, Q.Plath, K.David Geffen School of Medicine, University of California, Los Angeles, CA 90095, USA.Role of the murine reprogramming factors in the induction of pluripotencyCellCell364-771362AnimalsCell DifferentiationDNA-Binding Proteins/metabolismEmbryonic Stem Cells/*cytologyFibroblasts/cytologyMice/*metabolismNuclear Proteins/metabolism*Nuclear ReprogrammingPluripotent Stem Cells/*cytology2009Jan 2319167336http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=19167336[4] note that the changes associated with Oct4, Sox2 and Klf4 (full induction) are skewed towards transcriptional regulation. Transcriptional regulation is represented well in the PluriNetWork, and we expect to find startups here. Sridharan et al. also hypothesize that Nanog may be a key factor for full induction.
Figure SEQ Fig. \* ARABIC 11. Transition form piPS to iPS.
Indeed, by ExprEssence condensation of the PluriNetWork highlighting putative mechanisms of full induction (Figure 11), we observe Nanog-driven startup of pluripotency-related transcription factors, such as Esrrb, Sall4, Tbx3, Zfp42 and Zic3. Sridharan et al. also report induction of 21 genes (Figure 4D in their paper). Inspecting their list, of the five genes included in the PluriNetWork (Tcfcp2l1, Lefty2, Tcl1, Dppa4 and Klf2), startup of stimulation of Tcfcp2l1, Dppa4 and Klf2 is highlighted in the condensed network (Figure 11). Startup is below-threshold in case of Lefty2, and Tcl1 is only connected to genes with negligible change in expression value; it is not connected to the main pluripotency factors. Also, inhibitions around Nanog are shut down, including inhibitions by Cdx2 (trophectoderm marker) and Gata6 (primitive endoderm marker). Coincidentally, partial induction involves shutdown of inhibition by Cdx2 only (see above), whereas full induction involves shutdown of inhibitions by both Cdx2 and Gata6, reflecting the embryonal developmental timeline. Finally, Sox2 stimulation of Dppa4 starts up, concordant with Sridharan et al. However, even though Sox2 changes about as much as Nanog in its expression level, this is the only highlighted effect around the Sox2 gene; most of the putative action happens around Nanog, concordant with Sridharan et al.
Embryonic Stem Cell to Epiblast Stem Cell state
Transition from embryonic to epiblast stem cell state.
As described in the PluriNetWork paper, we used microarray data by Greber et al. ADDIN EN.CITE Greber201046464617Greber, B.Wu, G.Bernemann, C.Joo, J. Y.Han, D. W.Ko, K.Tapia, N.Sabour, D.Sterneckert, J.Tesar, P.Scholer, H. R.Department of Cell and Developmental Biology, Max Planck Institute for Molecular Biomedicine, Rontgenstrasse 20, D-48149 Munster, Germany.Conserved and divergent roles of FGF signaling in mouse epiblast stem cells and human embryonic stem cellsCell Stem CellCell Stem Cell215-2663AnimalsCell Culture TechniquesCell DifferentiationCell LineageCells, CulturedEmbryonic Stem Cells/cytology/*metabolismFemaleFibroblast Growth Factors/*metabolismGene Expression RegulationGerm Layers/cytology/*metabolismHumansMice*Signal TransductionSmad2 Protein/metabolismSmad3 Protein/metabolismStem Cells/*metabolism2010Mar 520207225http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=20207225[11] (GSE10017) to investigate the transition from the embryonic stem cell state to the epiblast stem cell state.
Input: Microarray data describing four different conditions as described below (see HYPERLINK \l "Table__9" Table 9).Analysis:ExprEssence highlighting of the top startups and shutdowns.Output: Hypotheses for (regulatory) mechanisms acting in the transition from the embryonic stem cell state to the epiblast stem cell state.Interpretation: Comparison of our results with the literature, and confirmatory experiments (see Figure 8 of the HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/PluriNetWork.pdf" PluriNetWork paper).
Table SEQ Table \* ARABIC 6. Workflow of the ExprEssence analysis of the ES / Epiblast transition.
ExprEssence is now used to derive the first condensed network, comparing (1) 12h PD LIF and (2) 12h PD Jaki. In the General ExprEssence Workflow HYPERLINK \l "GEW" above, the following data are provided, resulting in Figure 12B below; note that Figure 12 coincides with Figure 7 of the HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/PluriNetWork.pdf" PluriNetWork paper:
Figure
Steps
12BFile: HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/Epiblast.cys" Epiblast.cysAttribute:
mean valuesleft: 12h_PD+LIF_Signal
right: 12h_PD+JAKi_SignalAttribute:
variances (replicates)-
-Slider:
lower5%Slider:
upper95%
Table SEQ Table \* ARABIC 7. User input for reproduction of Figure 12B.
(The specific layout found in the figures below can be achieved by employing an organic layout with nodes of equal size, using the corresponding button in the ExprEssence tab, and manually moving the nodes.)
Figure SEQ Fig. \* ARABIC 12. Condensed network based on comparing (1) 12h PD LIF and (2) 12h PD Jaki, based on an older version of the PluriNetWork (panel A), and the version described in the paper (panel B).
In the PluriNetWork paper, we compared the resulting condensed network (Figure 12B) to a network condensed with exactly the same microarray data, but with an older version of the PluriNetWork (Figure 12A, taken from ADDIN EN.CITE Warsow16216216217Warsow, GregorGreber, BorisFalk, SteffiHarder, ClemensSiatkowski, MarcinSchordan, SandraSom, AnupEndlich, NicoleSchler, HansRepsilber, DirkEndlich, KarlhansFuellen, G.ExprEssence Revealing the essence of differential experimental data in the context of an interaction/regulation networkunder review[12]), and discussed some of the differences. Some more differences are apparent, and they are discussed here. In particular, as Klf4 is known to be upstream of Esrrb ADDIN EN.CITE Jiang200854545417Jiang, J.Chan, Y. S.Loh, Y. H.Cai, J.Tong, G. Q.Lim, C. A.Robson, P.Zhong, S.Ng, H. H.Gene Regulation Laboratory, Genome Institute of Singapore, 60 Biopolis Street, #02-01, Genome Building, Singapore 138672.A core Klf circuitry regulates self-renewal of embryonic stem cellsNat Cell BiolNat Cell Biol353-60103AnimalsApoptosisCell DifferentiationDNA-Binding Proteins/*physiologyEmbryonic Stem Cells/*cytologyEnhancer Elements, Genetic*Gene Expression RegulationGene Expression Regulation, DevelopmentalHomeodomain Proteins/*physiologyKruppel-Like Transcription Factors/*physiologyMiceModels, BiologicalTranscription Factors/metabolismTranscription, Genetic2008Mar18264089http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=18264089[13], we were puzzled that Klf4 is no longer featured in the condensed network based on the PluriNetWork we investigate here (Figure 12B). Moreover, we wondered why the presumed shutdown of Stat3-related signaling is not highlighted much stronger. Finally, we wondered why the shutdown of stimulation of Stat3 by Jak is not highlighted.
In Figure 12 (panels A and B), only the 5% most strongly differentially altered links of the PluriNetWork are displayed. For a threshold higher than 6.5%, the regulatory links between Klf4 and Esrrb appear, though (Figure 13A).
Figure SEQ Fig. \* ARABIC 13. Condensed network based on comparing conditions (1) 12h PD LIF and (2) 12h PD Jaki, using the original gene expression values (panel A) and log-transformed values (panel B).
In the General ExprEssence Workflow HYPERLINK \l "GEW" above, the following data must be provided, resulting in Figure 13 above. (Figure 13B is based on log-transformed data, to be discussed below.)
Figure
Steps
13A
13BFile: HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/Epiblast.cys" Epiblast.cys HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/Epiblast.cys" Epiblast.cysAttribute:
mean valuesleft: 12h_PD+LIF_Signal
right: 12h_PD+JAKi_Signalleft: X12h_PD.LIF_Signal
right: X12h_PD.JAKi_SignalAttribute:
variances (replicates)-
--
-Slider:
lower6.5%6.5%Slider:
upper93.5%93.5%
Table SEQ Table \* ARABIC 8. User input for reproduction of Figure 13.
In fact, while the ExprEssence link score measuring differential alteration of gene expression is only dependent on the change of the expression values of the two genes being linked, the addition of new nodes and links to the PluriNetWork modified the overall distribution of link scores so that the links between Klf4 and Esrrb are no longer below the 5% threshold, and they are only displayed if we visualize the 6.5% most differentially altered links. Closer inspection of Klf4, Klf2, Klf5, Esrrb and Nanog expression values and link scores (Table 9) reveals the reason underlying the fragility in observing Klf4 in our analysis of prominent mechanisms: In terms of the original value of the intensities in the data of GSE10017, Klf4 is expressed at a relatively low level, and its change (difference in intensities, not ratio of intensities) is low on this original scale.
Gene12h PD LIF12h PD JakiLink score to EsrrbAbsolute intensity valueLog-transformed intensity valueAbsolute intensity valueLog-transformed intensity valueAbsolute intensity valueLog-transformed intensity valueEsrrb5173.3512.344658.3812.19-1029.93-0.30Klf22705.7011.402379.0411.22-841.62-0.34Klf4306.388.2690.906.51-730.44-1.90Klf52647.4411.371893.1810.89-1269.22-0.64Nanog1410.0910.461022.7110.00-902.34-0.61
Table SEQ Table \* ARABIC 9. Selected absolute and log-transformed gene expression values and link scores for the ES to Epiblast transition. Absolute and log-transformed values of gene expression intensity for selected genes from GSE10017, for conditions (1) 12h PD LIF and (2) 12h PD Jaki, and corresponding link scores. The ExprEssence formula for the link score from a gene (Klf2, Klf4, Klf5, Nanog, Esrrb) to Esrrb takes the sum, over both genes, of the difference in gene expression between the two conditions under consideration (subtracting the value of condition 12h PD LIF from that of condition 12h PD Jaki). For example, in case of Klf4Esrrb, we obtain for original values (90.90-306.38)+(4658.38-5173.35)= -730.45 and for log2-transformed values we calculate (6.51-8.26)+(12.19-12.34)= -1.9. For Esrrb, the link scores to itself are derived from its self-stimulation link.
The original scale highlights differential alterations in expression on any level, in this case focusing on the cooperation of Klf5, Klf2 and Nanog. Klf4 acts at a much lower concentration (assuming that concentration correlates well with gene expression), and its prominent role is thus revealed if we transform the data to a logarithmic scale (Figure 13B). On the logarithmic scale, the cooperation of the other factors at higher expression/concentration is no longer prominent in our case, and Klf4 takes center stage in downregulating Esrrb. We also observe that Stat3 is now highlighted much stronger as a hub of a large number of shutdowns, and this holds true to for Klf4 as well. The shutdown of Stat3 stimulation by Jak is now highlighted as well, though it does not receive a high link score. The reason for this is simply that our highlighting is based on differences in gene expression, but changes in pathway activity (here in case of the Jak/Stat pathway) do not necessarily imply changes in gene expression.
The question arises whether logarithmic or original intensity data shall be used for including gene expression data into networks such as the PluriNetWork. As demonstrated by ADDIN EN.CITE Li200619319319317Li, W.Suh, Y. J.Zhang, J.Robert S Boas Center for Genomics and Human Genetics, Feinstein Institute for Medical Research, North Shore LIJ Health System, Manhasset, NY 11030, USA. wli@nslij-genetics.orgDoes logarithm transformation of microarray data affect ranking order of differentially expressed genes?Conf Proc IEEE Eng Med Biol SocConf Proc IEEE Eng Med Biol Soc6593-6SupplOligonucleotide Array Sequence Analysis/*methods*Signal Processing, Computer-Assisted200617959461http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=17959461[14], the logarithmic transformation may affect results on selecting differentially expressed genes. The study by Huang & Qu ADDIN EN.CITE Huang200615615615617Huang, S.Qu, Y.Statistics and Information Science, Lilly Corporate Center, Eli Lilly and Company, Indianapolis, IN 46285, USA. huang_shuguang@lilly.comThe loss in power when the test of differential expression is performed under a wrong scaleJ Comput BiolJ Comput Biol786-97133Databases, GeneticGene Expression Profiling/standards*Models, GeneticObserver Variation*Oligonucleotide Array Sequence Analysis/standardsReference Standards2006Apr16706725http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=16706725[15] states that without prior assumptions, there is no clear-cut recommendation, though. We believe that both kinds of data give complementary insight if network links are sorted by a measurement of importance such as the ExprEssence link score: the differential analysis of original data highlights changes across the full spectrum of intensities, and these can well be the cooperation of transcription factors all expressed on a medium level. Logarithmic data are better in the differential detection of small changes on a low expression level, often triggered by the high-affinity binding of a single transcription factor, exemplified by Klf4 and Stat3 in our example.
To further understand the relation of PluriNetWork data and expression data, we investigated two further conditions examined by ADDIN EN.CITE Greber201046464617Greber, B.Wu, G.Bernemann, C.Joo, J. Y.Han, D. W.Ko, K.Tapia, N.Sabour, D.Sterneckert, J.Tesar, P.Scholer, H. R.Department of Cell and Developmental Biology, Max Planck Institute for Molecular Biomedicine, Rontgenstrasse 20, D-48149 Munster, Germany.Conserved and divergent roles of FGF signaling in mouse epiblast stem cells and human embryonic stem cellsCell Stem CellCell Stem Cell215-2663AnimalsCell Culture TechniquesCell DifferentiationCell LineageCells, CulturedEmbryonic Stem Cells/cytology/*metabolismFemaleFibroblast Growth Factors/*metabolismGene Expression RegulationGerm Layers/cytology/*metabolismHumansMice*Signal TransductionSmad2 Protein/metabolismSmad3 Protein/metabolismStem Cells/*metabolism2010Mar 520207225http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=20207225[11]: (3) 12h FGF LIF and (4) 12h FGF Jaki, see Table 10.
Treatment conditionLIFLIF inhibition by Jaki(effect on) downstream targets
(in italics)Stat3
Klf4(
Stat3
(
Klf4FGF inhibition by PD
(1) 12h PD LIF
( ES state maintained
(2) 12h PD Jaki
( partial transition to Epiblast(
MEK/ERK
(
Klf2FGF
(3) 12h FGF LIF
( partial transition to Epiblast
(4) 12h FGF Jaki
( transition to EpiblastMEK/ERK
Klf2
Table SEQ Table \* ARABIC 10. Summary of the four different treatment conditions applied to mouse embryonic stem cells in ADDIN EN.CITE Greber201046464617Greber, B.Wu, G.Bernemann, C.Joo, J. Y.Han, D. W.Ko, K.Tapia, N.Sabour, D.Sterneckert, J.Tesar, P.Scholer, H. R.Department of Cell and Developmental Biology, Max Planck Institute for Molecular Biomedicine, Rontgenstrasse 20, D-48149 Munster, Germany.Conserved and divergent roles of FGF signaling in mouse epiblast stem cells and human embryonic stem cellsCell Stem CellCell Stem Cell215-2663AnimalsCell Culture TechniquesCell DifferentiationCell LineageCells, CulturedEmbryonic Stem Cells/cytology/*metabolismFemaleFibroblast Growth Factors/*metabolismGene Expression RegulationGerm Layers/cytology/*metabolismHumansMice*Signal TransductionSmad2 Protein/metabolismSmad3 Protein/metabolismStem Cells/*metabolism2010Mar 520207225http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=20207225[11].
In both cases, the FGF/MEK/ERK cascade was activated by supplementing the ES cell culture medium with FGF2 (and no inhibition of FGF/MEK/ERK signaling by PD (which is short for PD0325901) was done). Combined FGF/Jaki treatment in condition (4) promotes most effectively the transition to the epiblast state (indicated by dark grey background). As described, it is generally assumed that JAK inhibition triggers a transition signal via Stat3 and Klf4, while FGF stimulation shall de-repress Klf2 through activation of the MEK/ERK cascade. Thus, contrasting conditions (1) 12h PD LIF and (2) 12h PD Jaki, we observed a strong effect on Klf4 expression (Figure 13A and B). However, contrasting conditions (1) 12h PD LIF and (3) 12h FGF LIF, we would instead expect that ERK and Klf2 take center stage. In Figure 14A and B we find the most strongly differentially altered links for this scenario, using the absolute intensity and log-transformed data, respectively.
Figure SEQ Fig. \* ARABIC 14. Condensed network based on conditions (1) 12h PD LIF and (3) 12h FGF LIF, using the original gene expression values (panel A) and log-transformed values (panel B).
In the General ExprEssence Workflow HYPERLINK \l "GEW" above, the following data are provided, resulting in Figure 14A and B above.
Figure
Steps
14A
14BFile: HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/Epiblast.cys" Epiblast.cys HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/Epiblast.cys" Epiblast.cysAttribute:
mean valuesleft: 12h_PD+LIF_Signal
right: 12h_FGF2+LIF_Signalleft: X12h_PD.LIF_Signal
right: X12h_FGF2.LIF_SignalAttribute:
variances (replicates)-
--
-Slider:
lower6.5%6.5%Slider:
upper93.5%93.5%
Table SEQ Table \* ARABIC 11. User input for reproduction of Figure 14.
Using absolute intensities (Figure 14A), we indeed observe shutdown of three Klf2 links: its Nanog stimulation, Fgf5 inhibition and Nr5a2 (also known as LRH-1) interaction all go down. Interestingly, stimulations and inhibitions by and of Klf5 start up. The log-transformed data (Figure 14B) are highlighting a complementary hypothesis, namely the shutdown of Lefty1 activation, most prominently by Klf4 and Sox2. Lefty1 encodes an Activin antagonist. Since Activin signaling is known to be required for the maintenance of the epiblast cell state, this regulatory change might indicate a shift towards an epiblast-compatible expression of autocrine growth factors.
Finally, comparing condition (1) 12h PD LIF to the mixed case of condition (4) 12h FGF Jaki, displayed in Figure 15A and B, we find that both Klf2 and Klf4 (as well as Esrrb and Nanog) are at the center of shutdowns (original intensities, Figure 15A).
Figure SEQ Fig. \* ARABIC 15. Condensed network based on comparing conditions (1) 12h PD LIF and (4) 12h FGF Jaki, using the original gene expression values (panel A) and log-transformed values (panel B)
In the General ExprEssence Workflow HYPERLINK \l "GEW" above, the following data are provided, resulting in Figure 15A and B above.
Figure
Steps
15A
15BFile: HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/Epiblast.cys" Epiblast.cys HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/Epiblast.cys" Epiblast.cysAttribute:
mean valuesleft: 12h_PD+LIF_Signal
right: 12h_FGF2+Jaki_Signalleft: X12h_PD.LIF_Signal
right: X12h_FGF2.Jaki_SignalAttribute:
variances (replicates)-
--
-Slider:
lower6.5%6.5%Slider:
upper93.5%93.5%
Table SEQ Table \* ARABIC 12. User input for reproduction of Figure 15.
The complementary focus on small-intensity effects (logarithmic intensities, Figure 15A) highlights shutdowns around Klf4, which changes dramatically, but at a low absolute expression level. Moreover, we hypothesize a strong startup of inhibition of the protooncogene Fos. As in the case of inhibition of p53 ADDIN EN.CITE Krizhanovsky200920520520517Krizhanovsky, V.Lowe, S. W.Stem cells: The promises and perils of p53NatureNature1085-64607259AnimalsCell AgingCell DivisionCyclin-Dependent Kinase Inhibitor p16/deficiency/genetics/metabolismFibroblasts/cytology/metabolismHumansMiceNeoplasms/metabolism/pathology/therapy*Nuclear ReprogrammingPluripotent Stem Cells/*cytology/*metabolismTumor Suppressor Protein p53/deficiency/genetics/*metabolism2009Aug 2719713919http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=19713919[16], a pro-proliferative effect may be the result.
NodeColor Visualisation of Gene Expression
The NodeColor function of ExprEssence can be used to visualize gene expression values over many experiments for all genes in a network displayed using Cytoscape. You can select any number of attributes, where each attribute contains the expression values (floating point numbers) for a specific experimental condition. The default color scheme (green for low values, white for intermediate values, red for high values) can be adapted. As a result, each node is displayed using a pie-chart, where slices are color-coded based on gene expression (see Figure 16).
General workflow for the NodeColor function
Open Cytoscape and open a Cytoscape session file (cys file). If not already done, import your own high throughput data (as described HYPERLINK "http://cytoscape.wodaklab.org/wiki/Cytoscape_User_Manual/Attributes" \l "Import_Attribute_Table_Files" here). In the Network tab of the Control Panel on the left, select the network for which you wish to visualize gene expression values.
Use the left mouse button to click on the color nodes-button in the toolbar.
Select the attribute(s) to visualize in the pie chart, by marking them and pushing the +-button.
Press OK.
Optional: Redefine the color scheme by changing the threshold values.
Press OK.
For all nodes in the network, the selected attributes will be displayed using pie-charts.
Figure SEQ Fig. \* ARABIC 16. NodeColor visualization based on the data in HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/MEF_piPS_iPS.cys" MEF_piPS_iPS.cys, displaying a part of the PluriNetWork (left: GSE14012_MEF_mean; right: GSE14012_piPS_mean).
Visualization of Oct4 knockdown data using NodeColor.
Visualizing the loss of pluripotency in the PluriNetWork should identify the agonists and antagonists of this cellular state. Taking the microarray data of ADDIN EN.CITE Niwa200014614614617Niwa, H.Miyazaki, J.Smith, A. G.Centre for Genome Research, The University of Edinburgh, King's Buildings, Edinburgh, UK.Quantitative expression of Oct-3/4 defines differentiation, dedifferentiation or self-renewal of ES cellsNat GenetNat Genet372-6244AnimalsCell DifferentiationCell DivisionCell LineCell LineageClone Cells/cytology/metabolismDNA-Binding Proteins/*biosynthesis/geneticsDown-Regulation/genetics*Gene Expression Regulation, Developmental*Genes, RegulatorGenes, ReporterMiceOctamer Transcription Factor-3RNA, Messenger/biosynthesisStem Cells/*cytology/metabolismTranscription Factors/*biosynthesis/genetics/metabolismTranscription, Genetic/geneticsTransfectionUp-Regulation/genetics2000Apr10742100http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=10742100[17], reported in ADDIN EN.CITE ADDIN EN.CITE.DATA [18] (GSE10477), we visualized change of mouse ES cell gene expression after two days of Oct4 (Pou5f1) conditional knockout. As described by ADDIN EN.CITE ADDIN EN.CITE.DATA [18], the cells begin to exhibit trophectodermal morphology after two days.
For visualization, we utilized the NodeColor function of ExprEssence. The NodeColor function is described in the HYPERLINK \l "NodeColorWorkflow" ExprEssence section.
Figure SEQ Fig. \* ARABIC 17. Visualization of Oct4 knockdown data using NodeColor.
Reproduction of Figure 17.
Use the Cytoscape file HYPERLINK "http://www.ibima.med.uni-rostock.de/IBIMA/PluriNetWork/Oct4_ko.cys" Oct4_ko.cys.
Press the color nodes button.
Figure SEQ Fig. \* ARABIC 18. The color nodes button.
Select the attributes GSE10573_Oct4_KO_d0_to_d1 and GSE10573_Oct4_KO_d1_to_d2. Press OK.
Use the threshold values low= 0, mediate= 8, high= 16. Press OK.
Expression values are visualized by coloring the nodes following the heatmap metaphor, mapping low values to green color, and high values to red color. Highly divergent values can then be identified easily by inspecting color difference. Intermediate levels of expression are represented by color of less intensity.
In the PluriNetWork of Figure 17, the coloring scheme allows us to observe an inverse correlation of expression values: agonists of pluripotency, including Oct4 (Pou5f1), Sox2, Nanog, Esrrb, Nr5a2 (also known as LRH-1), Klf2, Klf4, Klf5 and Fbxo15 are downregulated after two days. In contrast, antagonists of pluripotency, such as Cdx2, Tcfap2c, Gata6, Dkk1 (Wnt inhibitor), Cdkn1a & Fos (inhibited by Oct4), are more expressed after two days of Oct4 KO.
References
ADDIN EN.REFLIST 1. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, et al. (2000) Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25: 25-29.
2. Camon E, Magrane M, Barrell D, Lee V, Dimmer E, et al. (2004) The Gene Ontology Annotation (GOA) Database: sharing knowledge in Uniprot with Gene Ontology. Nucleic Acids Res 32: D262-266.
3. Cline MS, Smoot M, Cerami E, Kuchinsky A, Landys N, et al. (2007) Integration of biological networks and gene expression data using Cytoscape. Nat Protoc 2: 2366-2382.
4. Sridharan R, Tchieu J, Mason MJ, Yachechko R, Kuoy E, et al. (2009) Role of the murine reprogramming factors in the induction of pluripotency. Cell 136: 364-377.
5. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, et al. (2003) Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4: 249-264.
6. Huangfu D, Osafune K, Maehr R, Guo W, Eijkelenboom A, et al. (2008) Induction of pluripotent stem cells from primary human fibroblasts with only Oct4 and Sox2. Nat Biotechnol 26: 1269-1275.
7. Pardo M, Lang B, Yu L, Prosser H, Bradley A, et al. (2010) An expanded Oct4 interaction network: implications for stem cell biology, development, and disease. Cell Stem Cell 6: 382-395.
8. van den Berg DL, Snoek T, Mullin NP, Yates A, Bezstarosti K, et al. (2010) An Oct4-centered protein interaction network in embryonic stem cells. Cell Stem Cell 6: 369-381.
9. Niwa H, Toyooka Y, Shimosato D, Strumpf D, Takahashi K, et al. (2005) Interaction between Oct3/4 and Cdx2 determines trophectoderm differentiation. Cell 123: 917-929.
10. Liao B, Jin Y (2010) Wwp2 mediates Oct4 ubiquitination and its own auto-ubiquitination in a dosage-dependent manner. Cell Res 20: 332-344.
11. Greber B, Wu G, Bernemann C, Joo JY, Han DW, et al. (2010) Conserved and divergent roles of FGF signaling in mouse epiblast stem cells and human embryonic stem cells. Cell Stem Cell 6: 215-226.
12. Warsow G, Greber B, Falk S, Harder C, Siatkowski M, et al. ExprEssence Revealing the essence of differential experimental data in the context of an interaction/regulation network. under review.
13. Jiang J, Chan YS, Loh YH, Cai J, Tong GQ, et al. (2008) A core Klf circuitry regulates self-renewal of embryonic stem cells. Nat Cell Biol 10: 353-360.
14. Li W, Suh YJ, Zhang J (2006) Does logarithm transformation of microarray data affect ranking order of differentially expressed genes? Conf Proc IEEE Eng Med Biol Soc Suppl: 6593-6596.
15. Huang S, Qu Y (2006) The loss in power when the test of differential expression is performed under a wrong scale. J Comput Biol 13: 786-797.
16. Krizhanovsky V, Lowe SW (2009) Stem cells: The promises and perils of p53. Nature 460: 1085-1086.
17. Niwa H, Miyazaki J, Smith AG (2000) Quantitative expression of Oct-3/4 defines differentiation, dedifferentiation or self-renewal of ES cells. Nat Genet 24: 372-376.
18. Endoh M, Endo TA, Endoh T, Fujimura Y, Ohara O, et al. (2008) Polycomb group proteins Ring1A/B are functionally linked to the core transcriptional regulatory circuitry to maintain ES cell identity. Development 135: 1513-1524.
About
Authors: Clemens Harder, Gregor Warsow, Georg Fuellen
Publications:
Contact: HYPERLINK "mailto:fuellen@uni-rostock.de" fuellen@uni-rostock.de
! f g ζΣzbJ/z/ 4h--z h#S 5B*CJ OJ QJ ^J aJ mH ph sH .h--z h#S 56CJ OJ QJ ^J aJ mH sH .h--z h#S 56CJ OJ QJ ^J aJ mH sH +h--z h#S 5CJ OJ QJ ^J aJ mH sH %hB) 5CJ OJ QJ ^J aJ mH sH %h+ 5CJ OJ QJ ^J aJ mH sH h+ h+ 0J mH sH h+ h+ mH sH j h+ U%h=% 5CJ OJ QJ ^J aJ mH sH +h+ h+ 5CJ OJ QJ ^J aJ mH sH @
T
7 e g + , = gd#S ^ gd#S
&F gd#S
&